Skip to main content

Full text of "Energy accounting as a policy analysis tool: prepared for the Subcommittee on Energy Research, Development, and Demonstration of the Committee on Science and Technology, U.S. House of Representatives, Ninety-fourth Congress, second session"

See other formats


[COMMITTEE PEINT] 



ENERGY ACCOUNTING AS A POLICY 
ANALYSIS TOOL 



PREPARED FOR THE 

SUBCOMMITTEE ON ENERGY RESEARCH, 
DEVELOPMENT, AND DEMONSTRATION 

OF THE 

COMMITTEE ON 

SCIENCE AND TECHNOLOGY 

U.S. HOUSE OF REPRESENTATIVES 

NINETY-FOURTH CONGRESS 

SECOND SESSION 

BY THE 

ENVIRONMENT AND NATURAL RESOURCES DIVISION 

CONGRESSIONAL RESEARCH SERVICE 

LIBRARY OF CONGRESS 

Serial CC 







.TUNE 1J>7« 






:<¥.>* 



7 



J.,,4'* 



Printed for the use of the Committee on Science and Technology 



[COMMITTEE PKINT] 



ENERGY ACCOUNTING AS A POLICY 
ANALYSIS TOOL 



PREPARED FOR THE 

SUBCOMMITTEE ON ENERGY RESEARCH, 
DEVELOPMENT, AND DEMONSTRATION 

OF THE 

COMMITTEE ON 

SCIENCE AND TECHNOLOGY 

U.S. HOUSE OF REPRESENTATIVES 

NINETY-FOURTH CONGRESS 

SECOND SESSION 

BY THE 

ENVIRONMENT AND NATURAL RESOURCES DIVISION 

CONGRESSIONAL RESEARCH SERVICE 

LIBRARY OF CONGRESS 

Serial CC 




JUNE 1976 



Printed for the use of the Committee on Science and Technology 



U.S. GOVERNMENT PRINTING OFFICE 
68-391 O WASHINGTON : 1976 



For sale by the Superintendent of Documents, U.S. Government Printing Office 
Washington, D.C. 20402 - Price $5.46 



COMMITTEE ON SCIENCE AND TECHNOLOGY 
OLIN B. TEAGUE, Texas, Chairman 



CHARLES A. MOSHER, Ohio 

ALPHONZO BELL, California 

JOHN JARMAN, Oklahoma 

JOHN W. WYDLER, New York 

LARRY WINN, Jr., Kansas 

LOUIS FREY, Jr., Florida 

BARRY M. GOLDWATER, Jr., California 

MARVIN L. ESCH, Michigan 

JOHN B. CONLAN, Arizona 

GARY A. MYERS, Pennsylvania 

DAVID F. EMERY, Maine 

LARRY PRESSLER, South Dakota 



KEN HECHLER, West Virginia 
THOMAS N. DOWNING, Virginia 
DON FUQUA, Florida 
JAMES W. SYMINGTON, Missouri 
WALTER FLOWERS, Alabama 
ROBERT A. ROE, New Jersey 
MIKE McCORMACK, Washington 
GEORGE E. BROWN, Jr., California 
DALE MILFORD, Texas 
RAY THORNTON, Arkansas 
JAMES H. SCHEUER, New York 
RICHARD L. OTTINGER, New York 
HENRY A. WAXMAN, California 
PHILIP H. HAYES, Indiana 
TOM HARKIN, Iowa 
JIM LLOYD, California 
JEROME A. AMBRO, New York 
CHRISTOPHER J. DODD, Connecticut 
MICHAEL T BLOUIN, Iowa 
TIM L. HALL, Illinois 
ROBERT (BOB) KRUEGER, Texas 
MARILYN LLOYD, Tennessee 
JAMES J. BLANCHARD, Michigan 
TIMOTHY E. WIRTH, Colorado 

John L. Swigert, Jr., Executive Director 
Harold A. Gould, Deputy Director 

Philip B. Yeager, Counsel 

Frank R. Hammill, Jr., Counsel 

James E. Wilson, Technical Consultant 

J. Thomas Ratchford, Science Consultant 

John D. Holmfeld, Science Consultant 

Ralph N. Read, Technical Consultant 

Robert C. Ketcham, Technical Consultant 

Regina A. Davis, Chief Clerk 
Michael A. Superata, Minority Counsel 



Subcommittee on Energy Research, Development, and Demonstration 



KEN HECHLER, West Virginia 
DON FUQUA, Florida 
JAMES W. SYMINGTON, Missouri 
GEORGE E. BROWN, Jr., California 
RAY THORNTON, Arkansas 
RICHARD L. OTTINGER, New York 
HENRY A. WAXMAN, California 
PHILIP H. HAYES, Indiana 
TOM HARKIN, Iowa 
JEROME A. AMBRO, New York 
CHRISTOPHER J. DODD, Connecticut 
ROBERT (BOB) KRUEGER, Texas 
MARILYN LLOYD, Tennessee 
JAMES J. BLANCHARD, Michigan 
TIMOTHY E. WIRTH, Colorado 



MIKE McCORMACK, Washington. Chairman 

BARRY M. GOLDWATER, Jr., California 

ALPHONZO BELL, California 

JOHN W. WYDLER. New York 

LARRY WINN, Jr., Kansas 

LOUIS FREY, JR., Florida 

MARVIN L. ESCH, Michigan 

JOHN B. CONLAN, Arizona 



(II) 



LETTER OF TRANSMITTAL 



House of Representatives, 
Committee on Science and Technology, 

Washington, D.C., Jime 8, 1976. 
Hon. Olin E. Teague, 

Chairman, Committee on Science and Technology, 
House of Representatives, Washington, D.C. 

Dear Mr. Chairman : Energy accounting, or energy analysis as it 
is sometimes called, is often cited as a basis for support of or objection 
to policy alternatives we consider in our authorizations of energy re- 
search, development and demonstration programs. In the Federal 
Non-Nuclear Energy Research and Development Act of 1974, Congress 
required that "the potential for production of net energy . . . shall 
be analyzed and considered in evaluating proposals" for alternative 
energy supply for technologies. 

The energy accounting literature is growing rapidly, and opinions 
as to its relevance to policy analysis vary widely. Accordingly, I have 
asked the Congressional Research Service to review this growing body 
of literature and assess its current relevance to our legislative needs. 
The CRS report, supplemented by representative writings in the field 
of energy accounting, concludes that the technique is not yet well 
enough developed to live up to many of the claims made for it. I believe 
that this judgment, which speaks primarily to the -question of current 
utility of the technique, will be of interest to the Committee in 
its deliberations on alternative energy supply and conservation 
technologies. 

Mike McCormack, 
Chairman. Subcommittee on Energy Research, 

Development and Demonstration. 

(in) 



Digitized by the Internet Archive 
in 2013 



http://archive.org/details/energyaccOOIibr 



LETTER OF SUBMITTAL 



The Library of Congress, 
Congressional Research Service, 

Washington, D.C., March 8, 1976. 
Hon. Mike McCormack, 

Chairman, Subcommittee on Energy Research, Development, and 

Demonstration, Committee on Science and Technology. U.S. 

House of Representatives, Washington, D.C. 

Dear Mr. Chairman : I am submitting herewith a report entitled 

"Energy Accounting as a Policy Analysis Tool" which was prepared at 

your request. The report describes the essential elements of Energy 

Accounting, traces its development over the past several years as an 

analytical technique, and measures its potential utility in policy 

analysis against its utility as demonstrated to date. 

The report, supplemented by a number of selected writings from the 
field, was prepared by David E. Gushee, Specialist in Environmental 
Policy in our Environment and Natural Resources Division. 

Norman Beckman, 
Acting Director, Congressional Research Service. 



(V) 



CONTENTS 



Page 

Letter of submittal in 

Letter of transmittal v 

Introduction 3 

What is Energy Accounting? 5 

Methodology 5 

Energy Flow Data 8 

Contemporary Analyses 8 

Conclusions 13 

APPENDIX 19 

(VII) 



ENERGY ACCOUNTING AS A POLICY ANALYSIS TOOL 

By David E. Gushee 

Specialist in Environmental Policy, Environment and Natural 
Resources Division 

March 8, 1976 

The Library of Congress 

4 

Congressional Research Service 



INTRODUCTION 



Energy Accounting 

The oil embargo of 1973-74 brought into sharp focus what is known 
and what is not known about energy flows in the U.S. A great deal is 
known on the supply side — how much oil, gas, coal, and uranium ore 
are produced, where it comes from, how it is transported, and where 
it is used. Less is known on the demand side — how consumption varies 
with capacity utilization, how much is used for heat, for processes, 
or for raw materials. And still less is known about interfuel substi- 
tutability, potential for conservation, and impact of changes in con- 
sumption patterns on raw energy supply needs. 

These information gaps are not surprising. So long as energy sup- 
plies were very large compared to demand, energy flows have been 
accounted for as components of dollar flows and flows of raw mate- 
rials and products. Energy inputs were not constraining; there was 
no need to identify them separately and treat them as independent 
variables, except as cost elements in financial accounting systems. 

The situation has now changed. Not only is energy more expensive 
(which means it has become worthy of special consideration in cost 
control), but its availability is no longer assured. Where it comes 
from, how it is used, whether it can be substituted for, how much is 
wasted, all become significant policy questions not only for individ- 
uals and companies but for the Nation as a whole. 

Predicting direct and indirect effects of changes in energy supply 
and usage patterns is also critical. Does saving energy in one walk of 
life lead to a savings in overall energy demand or does it, as a result 
of the complexity and perversity of modern life, lead instead to a net 
increase in overall energy demand? If it does conserve, who benefits, 
how, and where ? Who loses ? 

For some years before the Arab oil embargo of 1973-74, a number 
of close observers of the energy scene were predicting that trends in 
energy consumption were cause for long term concern. Not only had 
geologists such as M. King Hubbert been predicting that U.S. oil and 
gas reserves would soon begin to deplete, but foreign policy experts, 
among others, were predicting that the U.S. economy, fueled as it is 
primarily by oil and gas, would soon be in a position of import de- 
pendence. Studies were made and warnings were raised, but little was 
done, as we all know. 

A small coterie of physical scientists were, however, trying in those 
years to map energy flows through the U.S. economy to see how plen- 
tiful fuels such as coal might be substituted for depleting fuels like 
oil and gas, where energy might be conserved, and how the American 
lifestyle might be modified to reduce its enormous appetite for energy. 
Ecologists were concerned about the impact on the ecosphere of con- 
tinued growth in energy demand. 

(3) 



They had not progressed very far at the time the embargo hit. They 
are not very much farther ahead now. But their work has attracted a 
great deal of attention, in part because of their claims for its values 
and in part because of the failures of economic tools to provide ade- 
quate insight into the consequences of economic and energy policy 
alternatives proposed in response to the embargo, the subsequent 
increases in world energy prices, and the slowdowns in the econo- 
mies of most of the oil-consuming nations. 

Mapping energy flows is now on the threshold of becoming an in- 
dependent discipline. It is called "energy accounting" by some, "en- 
ergy analysis" by others. Other names have also surfaced from time 
to time. Its purpose is to model energy flows in a manner analogous 
to the way that economics models money flows. Some of its more 
extreme proponents claim that energy is a more basic unit of value 
than money and that energy accounting should replace money ac- 
counting. 1 Most energy analysts, however, claim only that energy 
accounting can provide insights that economic analysis cannot and 
that, were its methodology and data base improved, it would provide 
a useful, if not indispensable, additional tool for predictive (and 
therefore for policy) purposes. 

Some skeptics claim that energy accounting is nothing more than the 
traditional chemical engineer's energy balance and that, when applied 
to policy problems, it reveals nothing not already known through 
economic analysis. 

As is the case with most new analytical techniques, the truth will 
eventually become clear, and the technique will take its rightful place. 
At present, energy accounting appears to be of potential value in 
evaluation of alternative energy supply technologies and in identifi- 
cation and ranking in priority of energy conservation opportunities. 
Whether it will ever become useful in improving economic forecasts, 
energy demand forecasts, or impact analysis of policy options such 
as increased recycling of resources or energy tax proposals is not yet 
clear. 



1 Hannon, Bruce M., An Energy Standard of Value. Annals of the American Academy of 
Political and Social Sciences, No. 410, 1973, pp. 139-53. Reprinted in Appendix, page 27. 



What Is Energy Accounting? 

Energy accounting is tracing the flows of energy into, through, and 
out of systems. A system can be defined at will : anything from a car 
to the U.S. economy as a whole, or to the economies of a number of 
countries or the world at large. The simpler and more specific the 
system, the more precise the analysis can be. As systems under analysis 
become more complex, generalized data, if available, must be used 
to reduce the computation volume ; if generalized data are not avail- 
able, they must be estimated from indirect sources such as economic 
data and materials flows. Resorting to generalized data or to esti- 
mated data introduces errors of unknown magnitudes and causes 
loss of detail. 

The type of analysis used and the definition of the system to be 
analyzed depend on the purpose of the analysis. As there are many 
ways to analyze and many things to be analyzed for, there are sig- 
nificant variations in basic assumptions, methodologies, and data 
estimation techniques in the studies carried out to date. These varia- 
tions make comparisons between studies difficult at best and often 
lead analysts to different conclusions. 

These variations are not unusual for a new discipline. As interest 
has mounted and new workers have entered the field, the need for 
some standardization has become recognized. Two international meet- 
ings have been held with this in mind — in August 1974 in Gulds- 
medshyttan, Sweden, and in June 1975 in Lidingo, Sweden. Both 
were held under the auspices of the International Federation of In- 
stitutes for Advanced Study. The 1974 meeting focused on the 
mechanics of energy accounting, including terminology, units of 
measurement, system definition, and format for reporting results. 
The 1975 meeting focussed on the relationship between energy analy- 
sis and economic analysis, seeking to find the aspects of each of po- 
tential value to the other. Reports of the two meetings are reproduced 
in the Appendix. 

These two meetings have helped to reduce the variations from study 
to study and have begun to build a bridge between energy accounting 
and economics. But much remains to be done before a body of agreed- 
upon analytical results becomes available and usable for further 
analyses. 

Given this background, energy accounting is developing on three 
broad fronts : 

Methodology 
Energy Flow Data 
Contemporary analyses 

I. METHODOLOGY 

There are two basic methodological approaches to energy account- 
ing: (a) process analysis, which treats individual processes or groups 

(5) 



6 

of processes, using either directly measured data or industry aver- 
ages; and (b) input/output analysis, which uses economic data on 
inter-industry transfers of goods and services transformed to esti- 
mates of associated energy transfers. 

A. Process Analysis 

In process analysis, the analyst makes what engineers call a "mate- 
rial balance" and an "energy balance." He makes a block diagram to 
include each process step and identifies the quantities and physical 
and thermodynamic states of materials and energy flowing in and 
out plus all internal transformations and flows. Where actual meas- 
ured values are available, those are used ; in their absence, the analyst 
makes estimates based on his knowledge of physical and chemical 
principles. 

Process analysis has the advantage of giving detailed insight into 
the process and how it varies as conditions change. The analyst can 
derive averaged information and incremental information — how unit- 
consumptions change as production is increased or decreased, for 
example. 

Process analysis, although it may be new to the policy arena, is a 
classic tool of the industrial process engineer, particularly in the 
chemical process industries such as iron and steel, petrochemical, 
metallurgical, and aluminum. There are three main differences, how- 
ever, between process analysis as practiced in industry and as prac- 
ticed for policy work. First, the process engineer has available, or can 
estimate from chemical and physical principles, all the numbers nec- 
essary to develop a complete energy picture of the process. Secondly, 
he is not interested in what went on before his process inputs reached 
him nor in what happens after his products and wastes leave his plant 
boundaries. Thirdly, he is not privy to how his energy balance com- 
pares to those of other processors in other companies at other 
locations. 

The first point — complete process knowledge — means that the in- 
dustrial engineer can identify energy conservation opportunities in his 
plant, can calculate the effects on energy consumption caured by 
changes in operating rate or process conditions, and can calculate the 
effects of process redesign or raw material changes. 

The second point — process isolation — means that he is less concerned 
with the impact on upstream or downstream energy flows that might 
be caused by changes he might make within his plant. A change from 
gas to coal for process heat, for example, would change his flowsheet 
and economics. What that would do to the railroad traffic or inter- 
state gas flow are not taken into account in his deliberations. 

The third point — competitive isolation — means that no individual 
in an industry has complete knowledge of the whole industry and thus 
cannot make an accurate process analysis for the industry as a whole. 
This is particularly important with respect to changes in operating 
rates as economic activity changes and to policy changes in fuel prices 
or availabilities. 

In sum, process analysis fits directly into energy conservation actions 
by industry on a case by case basis but is not directly suited for broader 
policy questions such as externally imposed constraints or require- 
ments. In order for it to provide meaningful predictions of impacts 



from such external impositions as fuel-switching requirements or 
mandatory conservation goals, the detailed data and analyses made 
on specific processes would have to be made available to some central 
entity where the impact on the whole could be calculated from the 
sums of impacts on the individuals. To date, only one industry — iron 
and steel — has been able to develop such a central capability, at 
Arthur D. Little, Inc., 2 in Cambridge, Mass. The data and mathe- 
matical model are held in tight secrecy to protect the proprietary 
individual company data, and not all the process units in the industry 
are represented, although the majority (130 of 139 plants of members 
of American Iron and Steel Institute, representing 99% of total AISI 
production) are. 

B. Input/ Output Analysis 

Input/output analysis, based on economic data on inter-industry 
transfers of goods and services, starts at the industry level and can be 
aggregated upward to groups of industries or combinations of indus- 
tries and sectors (such as transportation). By its very nature, input/ 
output cannot be disaggregated downward — that is, it cannot provide 
an accurate picture or predictive capacity for impacts of external 
changes within an industry. 

An input/output table shows the dollar flows from one industry 
into others — where iron and steel go, for example. Taken the other 
way, it shows where an industry, such as iron and steel, gets its inputs 
and how much it pays for them. It does not measure energy transfers, 
although it does measure fuel and electricity inputs as dollar flows. 
Nor does it show how the inputs are used — as feedstocks or process 
heat, for example, nor how they are distributed among the various 
functions within the industry. 

Further, input/output tabulations become available only some years 
after the fact. The latest complete set for the U.S., for example, is 
based on 1967 figures. To add further difficulty, the economic data 
must be converted to energy units on the basis of some average price. 
Where a significant portion of the energy flows are under vertically- 
integrated control, or prices vary widely as a function of rapidly- 
varying over and undersupply or seasonal demand, or deliveries are 
under long term price contracts whose terms are not known, the use 
of average price introduces an unknown error in energy flows and can 
misrepresent the probable impact of external change such as spot 
market price changes. 

Despite these shortcomings, input/output analysis does give num- 
bers for energy flows through economic sectors, and some of these 
numbers for some sectors are chimed by energy analysts to have 
policy value (for some purposes). The main value is apparently their 
reflection that every product or service represents an overall energy 
flow in the economy significantly larger than the energy directly 
required to make it, and that "gross energy requirement" is different 
for different types of products. A particularly promising result from 
such an analysis, for example, is that making polyethylene in the U.S. 
has a much greater gross energy requirement than making the same 



3 Steel and the Environment : A Cost Impact Analysis. Arthur D. Little, Inc.. Report to 
the American Iron and Steel Institute. May 1975. 



8 

material in the Netherlands, as a result of differing raw materials in 
the two countries. 3 Whether this result is right and if right whether 
it signals the need for some sort of policy initiative are not yet clear. 

Nonetheless, the main advantage of input/output analysis is the rela- 
tive ease of use and the opportunity to use the same set of starting data : 
the Federal input/output statistics. 

Articles describing and evaluating both process analysis and input/ 
output analysis are reprinted in the Appendix. In the U.S., K. Stephen 
Berry and Thomas V. Long and coworkers, University of Chicago, 
are the leading developers and practitioners of process analysis, while 
Bruce Hannon, Robert Herendeen, and coworkers Center for Ad- 
vanced Computation, University of Illinois, play the same role for 
input-output analysis. 

II. ENERGY FLOW DATA 

To date, there is no centralized collection of energy flow data. There 
are many data collections which include relevant, though partial, 
data — so many in fact that a recent Federal Energy Administration 
report 4 summarizing Federal energy information sources is more than 
two inches thick. Data are generated on fuel supply and transportation 
activities in great detail on a continuous basis. Fuel flows into indi- 
vidual industrial categories are also available. Fuel use patterns within 
industries and other consuming sectors are, however, measured only 
sporadically and vary in comprehensiveness of coverage and com- 
patibility with other energy flow data. 

Technical and trade publications and a number of handbooks and 
manuals also include energy data. These tend to be "typical" plant 
data representing an idealized process or plant or highly specific tc 
an individual set of circumstances. 

This mixed bag of data resources is very much like that for eco- 
nomics. Some activities are well documented; others are treated pri- 
marily only on an aggregated basis. There is a major gulf between the 
"micro" level and the "macro" level (specific systems vs. aggregated 
systems) which can be bridged only by making a number of assump- 
tions subject to great inherent uncertainty. 

III. CONTEMPORARY ANALYSES 

The energy accounting literature is expanding rapidly, as funding 
increases and new workers enter the field. The studies published to 
date fall into one or more of the three following categories : 

(a) Net energy analysis of energy supply systems. 

(b) Gross energy requirements of economic sector activities. 

(c) Energy impacts of price, supply, and technology alterna- 
tives and vice versa. 

A . Net Energy Analysis of Energy Supply Systems 

Net energy analysis applied to energy supply systems is designed 
to identify how much energy is used up in the process of extracting 



3 R. S. Berry. T. V. Long, and H. Makino. An International Comparison of Polymers and 
Their Alternatives. Energy Policy. Vol. 3. June 1975. pp. 144-155. Reprinted in Appendix 

4 Energy Information in the Federal Government. Federal Energy Administration NTIS 
No. PB. 246703. 1975. 



9 

an energy resource such as coal or uranium, transforming it to usable 
forms, and delivering it to the points of use. Its rise in popularity and 
potential value is associated with the current ongoing debate on the 
finiteness of non-renewable resources, ultimate "limits to growth," and 
the probable requirement to exploit resources that are progressively 
harder to get at and leaner in usable energy content. 

Although trends toward leaner and leaner ores have been recognized 
for generations, net energy yield as a rallying cry first surfaced about 
five years ago when E. J. Hoffman of the University of Wyoming 
calculated 5 that "the net energy realized by nuclear fission may be 
nearly a full order of magnitude (a factor of ten) below that pre- 
dicted. In other words, instead of plant or thermal efficiencies of 30%, 
the net plant efficiency realized — based on the ideal value for the fuel — 
may run as low as 3%." 

This calculation was admittedly imprecise, based as it was on a 
number of necessary assumptions about the shape of the ultimate 
nuclear fuel cycle and waste management procedures. Nonetheless, it 
was seized upon by opponents of nuclear power as an additional argu- 
ment in their favor. 

At about the same time, Howard T. Odum, an ecologist at University 
of Florida, was incorporating the net energy concept into a general 
concern over the ability of the world's ecosphere to absorb the ever- 
increasing heat loads of industrial and population growth. Although 
he had published a highly technical book on the subject, 6 his ideas 
really took hold only after he published a simplified, semitechnical 
article in 1973. 7 

Since 1973, the net energy concept has been developing along two 
separate paths : net energy yield of nuclear power as part of the debate 
on the future role of nuclear power in the U.S. and other national 
economies; and net energy yields of energy supply systems generally, 
but particularly with respect to alternative energy sources such as 
synthetic fuels, solar energy, and oil shale. 

On nuclear power, the debate in print started out questioning the 
energy yield per nuclear power plant. As analytical sophistication in- 
creased, analytical results began to show incontrovertibly that a new 
nuclear plant would pay back its original energy investment rapidly 
compared to its life expectancy even given conservative assumptions 
about energy needs associated with fuel reprocessing, waste manage- 
ment, and plant decommissioning. The latest studies now are converg- 
ing on each other in their estimates of net energy yield, showing that 
the initial energy investment is repaid within two years or less of plant 
start-up. These plants are expected to operate for at least 30 years 
thereafter contributing "net energy." 

As the original question of net energy yield for fission reactors was 
being shown to be unimportant, the focus of the debate shifted to the 
impact on overall energy demand as nuclear plant investment es- 
calated. Concern was expressed that, far from being an overall energy 
conserver, a burgeoning nuclear program would sharply increase 
overall demand for fossil fuels and, if rapid enough, would always 
have that effect. Later studies, however, have shown that a rate of 



5 Hoffman, E. J., Overall Efficiencies of Nuclear Power, December 1071, unpublished. 
a Odum, H. T., Environment. Power and Society, John Wilev, 1972. 

7 Odum. H. T., Energy, Ecology and Economics. Ambio, Vol. 2. No. 6, 1973, pp. 220-5. 
Reprinted in the Appendix, page 19. 



10 

growth of nuclear power having that effect would be insupportable on 
other grounds — either overall growth rate of electricity demand or 
constraints on manpower or materials. This debate, reprinted in the 
Appendix, appears to be cooling in significance as the planned rate 
of future construction of nuclear power plants slows down. 

On the more general question of net energy yields of a range of 
alternative energy supply systems, after an initial period of expan- 
siveness, claims for the value of net energy analyses are becoming 
more realistic and hedged with qualifications. One of the more recent 
reports 8 is typical : "In calculating the amount of energy it takes to 
get energy, we have examined each major step in the pathway from 
extraction to conversion to fuel as well as to electricity with some trans- 
mission losses factored in. A brief thumbing through the report will 
indicate how detailed the analysis has become and how inclusive it 
must be to accurately reflect energy inputs and outputs. ... (T)he key 
feature is that the method does not rely on one number to express the 
variety of issues associated with energy analysis. There is no 'net' 
energy calculation in the sense that we can provide a single number 
for policy making. Hopefully, we run a lesser risk that the complexity 
will result in confusion, than simplification will result in misunder- 
standing" (page 5). 

Although the complexity and lack of a single number do not 
necessarily negate the potential value of net energy analysis, they do 
demonstrate that the technique will not supersede other evaluative 
methods such as economic analysis and subjective judgment. The role 
it will play remains unclear, even though Congress has legislated 9 
that net energy analyses will be required components of evaluations 
of proposed energy supply alternatives : 

"... the comprehensive program in research, development, and demon- 
stration required by this Act shall be designed and executed according 
to the following principles: ... (5) the potential for production of 
net energy by the proposed technology at the state of commercial 
application shall be analyzed and considered in evaluating proposals." 

The conference report on S. 1283, 10 which became Public Law 93-577, 
goes on to explain : "The intent of (this) principle ... is that in the 
assignment of priorities for Federal encouragement of commercial 
applications of new energy technologies, consideration should be given 
to the net, as opposed to the gross, energy yield. The processes and 
facilities necessary to produce energy also consume energy, and in 
the case of certain technologies, this consumption may account for a 
substantial portion of the potential yield of the energy resource .... 
(In) the early research or development phases of new technologies, 
the projected applications may even involve a net loss of energy. This 
principle is not intended in any way to deter such research or to 
deter the demonstration of new technologies which are not energy 
efficient or cost effective in the early stages of development." 

The Energy Research and Development Administration (ERDA), 
the agency responsible for implementing Public Law 93-577, is moving 
cautiously in translating this requirement for net energy analyses into 

8 A Study to Develop Energy Estimates of Merit for Selected Fuel Technologies. J>evelop- 
ment Sciences. Inc., Contract 14-01-0001-2141 DSI 038. to U.S. Department of the Interior. 

» Public Law 93-577, The Federal Non-Nuclear Energy Research and Development Act of 
1974. Section 5(a) (5). « -™ « ^ *„ +*** 

10 House Report 93-1563, Conference Report on S. 1283, December 11, 1974. 



11 

action. It has assigned the responsibility to a newly established position 
of Assistant Director for Systems Analysis in the Office of Assistant 
Administrator for Planning and Analysis. The methodology to be 
used is still being worked out. In addition to the Development Sciences 
study, a number of others funded prior to the formation of ERDA 
have been made. Two of the more comprehensive of these "• 12 show 
some of the difficulties that this new technique faces in its development 
phase. 

To supplement the insights gained by these studies, ERDA, with 
assistance from the National Science Foundation which had been 
supporting methodological studies for several years, convened a work- 
shop of leading theoreticians and practitioners of energy analysis. 13 
The specialists in attendance were unable to hammer out agreement 
among themselves; since then, ERDA has been continuing its de- 
liberations on methodologies and is currently not expecting to finalize 
its approach for many months. 

B. Gross Energy Requirements 

Despite the limitations imposed by weaknesses and gaps in present 
analytical techniques and available data, energy analysts continue to 
apply both process analysis and input/output analysis to national 
economies and to various economic sectors. Although their results are 
subject to uncertainties, as indicated earlier, these analysts are be- 
ginning to make comparisons between countries and to generate num- 
bers for goods and services that represent the relative overall energy 
intensities for the activities going into their manufacture or their de- 
livery of service. Several such analyses are reprinted in the Appendix. 

It is difficult to find any unique policy value in these studies. One 
of the basic purposes of making them has been to identify situations 
where economic analysis does not adequately reveal the energy inten- 
siveness of an economic sector and therefore does not adequately 
predict the energy impacts of economic changes or the reverse. An 
oft-quoted statistic, for example, is the fact that the energy con- 
sumption per capita in the U.S. is twice as great as that for Sweden. 
Switzerland, or some other country with equivalent or nearly equivalent 
gross national product per capita. 

Energy analyses show significant differences between countries and 
between plants within a country in gross energy requirements for 
steel production, transportation, plastic production, and other goods 
and services. But the analyses do not yet show how much of these 
differences are real, and therefore the results of differences in tech- 
nological approach and energy costs, and how much might be the 
consequence of, for example foreign trade, which is a much larger 
factor in European economies than in the U.S. economy. 

Nor is it possible for energy analysts to calculate the consequences 
of systemic changes, such as switching from gas to coal or increasing 
the proportion of recycled materials. Nor is it possible for the analysts 
to show that their analyses have yet, in fact, uncovered situations where 
the energy analysis points to policy options that differ markedly from 
those indicated by traditional economic analysis. 

11 Energy Alternatives : A Comparative Analysis. Science and Public Policy Program, 
University of Oklahoma, May 1975. 

12 Environmental Impacts Efficiency and Cost of Energy Supply and End Use. Hittman 
Associates, January 1975, to Council on Environmental Quality and other agencies. 

13 Net Energy Analysis Workshop. August 25-28. 1975, sponsored by National Science 
Foundation. 



CONCLUSIONS 

Energy accounting is being offered as a policy analysis tool able to 
provide new or more defensible insights than have heretofore been 
available into the use of energy in the United States and other social 
and economic systems. Its proponents suggest that it can supple- 
ment, if not replace, existing analytical tools such as economic analysis 
and provide insurance against mistakes that might occur as a result of 
weaknesses in economics such as basic assumptions of input substituta- 
bility, the behavior of a free market, and the discount rate. 

Economic analysis, despite its long history and acceptance as a 
scholarly discipline and a contributor of insights, has many weak- 
nesses and limitations, as events of recent years demonstrate and as 
economists admit. It is not surprising, therefore, that energy account- 
ing, as a new methodology still seeking acceptance as a discipline, has 
many weaknesses and limitations and suffers the slings and arrows of 
many critics. 

On the record as exemplified by the articles included in the Appen- 
dix, one must qonclude that energy accounting has only limited 
policy value in its current state of development. In the short term, 
its greatest potential would appear to lie in identifying energy con- 
servation opportunities on a case by case basis and in providing an 
energy efficiency criterion for use in evaluating new technologies for 
energy production. 

Whether energy accounting will ever be able to supplement economic 
analysis as a tool to predict impacts of alternative public policy 
options is not yet clear. There is some possibility that its insights 
can enrich the economists' mathematical treatment of the form value 
of energy (variations in potential utility of the energy as a function 
of its physical and thermodynamic properties) and the applicability 
of the discount rate when applied to non-renewable resources. 14 
Neither of these enrichments has yet occurred in economic analysis, 
and economists are working on the same problems from other angles 
as well as from inputs provided by energy accounting. 

In sum, in its present state of development, energy accounting is 
worth following for its possible future value but appears to be of very 
limited value for current use. 



14 Talbot Page, Economics of Recycling, in Senate Public Works Committee Print Resource 
Conservation, Resource Recovery, and Solid Waste Management. Serial No. 93-12, 1973. 
Page discusses the economic issues associated with the conservation criterion, the factor 
through which a finite, depleting, nonrenewable resource is or might be given a current 
economic value. 

(13) 



APPENDIX 
I. Spreading Awareness: 

1. "Energy, Ecology, and Economics." Howard T. Odum, P a «" 

Ambio, v. 2, No. 6, 1973: pp. 220-227 19 

2. "An Energy Standard of Value." Bruce M. Hannon, Annals 

of the American Academy of Political and Social Science, 

v. 410, 1973; pp. 139-153 27 

3. "It Takes Energy to Produce Energy: The Net's the Thing." 

Edward Flattau and Jeff Stansbury, The Washington 
Monthly, March 1974; pp. 20-26 37 

4. "Systems of Energy and the Energy of Systems." Thomas A. 

Robertson, Sierra Club Bulletin, v. 60, March 1975; pp. 
20-23 41 

5. "The Old Economics Has Failed: A New System is Needed to 

Find the True Cost of Energy." Wade Rowland, Science 
Forum, v. 8, August 1975; pp. 3-6 46 

6. "It Takes Energy To Get Energy; the Law of Diminishing 

Returns is in Effect." Wilson Clark, Smithsonian, v. 5, 
December 1974; pp. 84-90 50 

7. "Energy Analysis and Public Policy." Martha W. Gilliland, 

Science, v. 189, September 26, 1975; pp. 1051-1056 55 

8. "Net Energy Analysis Can be Illuminating." Rice Odell, 

editor, Conservation Foundation Letter, October 1974 65 

II. Critics Begin to Surface: 

9. "Energy Accounting vs. the Market." Based on a paper by 

Joel Darmstadter, Resources, No. 50, October 1975; pp. 4-5_ 75 

10. "The Economics of Energy Analysis." Michael Webb and 

David Pearce, Energy Policy, v. 3, December 1975; pp. 
318-331 76 

11. "Net Energy Analysis — Is It Any Use?" Gerald Leach, Energv 

Policy, v. 3, December 1975 ; pp. 332-344 87 

III. Analytical Methodology: 

12. "Use of Input/Output Analysis to Determine The Energy 

Cost of Goods and Services." Robert A. Herendeen, in 
Energy Demand, Conservation, and Institutional Pro- 
blems, edited by Michael S. Macrakis, MIT Press, 1974__ 101 

13. "Energy costs: a review of methods." P. F. Chapman, Energy 

Policy, v. 2, June 1974; pp. 91-103 111 

14. Energy Analysis Workshop on Methodology and Conven- 

tions, Guldsmedshyttan, Sweden, August 1974; held under 
the auspices of the International Federation of Institutes 
for Advanced Study. Report No. 6 125 

15. Workshop on Energy Analysis and Economics, Lidingo, 

Sweden, June 1975; under the auspices of the International 
Federation of Institutes for Advanced Study. Report No. 9_ 214 

16. "Thermodynamics and Energy Accountancy in Industrial 

Processes." C. Cozzi, Energy Sources, v. 2, No. 2, 1975; 

pp. 165-178 325 

IV. The Nuclear Power Debate : 

17. "Energy Inputs and Outputs for Nuclear Power Stations." 

P. F. Chapman and W. D. Mortimer, Energy Research 
Group, Open University, Milton Keynes, Report ERG 
005, revised December 1974 335 

18. "Dynamic Energy Analysis and Nuclear Power." John H. 

Price, Friends of the Earth Ltd. (for Earth Resources 
Research Ltd.), December 1974 412 

19. "Nuclear Energy Balances in a World With Ceilings." Gerald 

Leach, International Institute for Environment and De- 
velopment, preliminary paper December 1974; unpub- 
lished 446 

(15) 



16 

IV. The Nuclear Power Debate — Continued Page 

20. "Energy analysis of nuclear power." John Wright and John 

Syrett, New Scientist, v. 65, January 9, 1975; pp. 66-67__ 473 

21. "Nuclear Power's Contribution to Energy Growth." W. Ken- 

neth Davis, paper presented at the Atomic Industrial 
Forum Conference on Accelerating Nuclear Power Plant 
Construction, March 3, 1975; New Orleans 475 

22. "The Net Energy from Nuclear Reactors." F. von Hippel, 

M. Fels and H. Krungmann. FAS Professional Bulletin, 

v. 3, April 1975; pp. 6-7 492 

23. "Energy Accounting and Nuclear Power." L. G. Brookes, 

Atom, v. 227, September 1975 ; pp. 164-168 494 

24. "Energy analysis of nuclear power: For and against nuclear 

power." Peter Chapman; "The growth of a myth." Len 
Brookes, New Scientist, v. 65, October 1975; pp. 142-147. 499 

25. Net Energy from Nuclear Power. R. M. Rotty, A. M. Perry, 

D. B. Reister, Institute for Energy Analysis Report IEA- 

75-3 (Summary and Conclusions) 503 

26. "Energy analysis of nuclear power stations." Peter W. 

Chapman, Energy Policy, v. 3, December 1975; pp. 
285-297 507 

27. "Nuclear power and oil imports: a look at the energy bal- 

ance." J. H. Hollomon, B. Raz, R. Triitel, Energy Policy, 

v. 3, December 1975; pp. 299-305 520 

28. "Energy analysis of a power generating system." K. M. Hill 

and F. J. Walford, Energy Policy, v. 3, December 1975; 

pp. 306-317 527 

V. Net Energy Yield of New Energy Supply Systems: 

29. Energy Study. Summary chapter, interim report, Office of 

Energy Research and Planning, Office of the Governor, 
State of Oregon. July 1974 541 

30. "The energy cost of fuels." P.F. Chapman, G. Leach, M. 

Slesser, Energy Policy, v. 2, September 1974; pp. 231-243. 544 

31. "Procedures for comparing the energy efficiencies of energy 

alternatives." Chapter 15, in Energy Alternatives: A 
Comparative Analysis. The Science and Public Policy 
Program, University of Oklahoma, May 1975 557 

32. "A Study to Develop Energy Estimates of Merit for Selected 

Fuel Technologies," summary chapter, Development 

Sciences, Inc., September 1975 575 

VI. Applications of Energy Analysis to National Economies and to 
Economic Sectors: 

33. "Total Energy Demand for Automobiles." Eric Hirst and 

Robert Herendeen. Paper 730065 delivered at the Inter- 
national Automotive Engineering Congress, Society of 
Automotive Engineers, January 1973 589 

34. "The Energy Cost of Automobiles." R.S. Berry and M.F. 

Fels; Science and Public Affairs, v. 29, December 1973; 

pp. 11-17 and 58-60 595 

35. "Energy Thrift in Packaging and Marketing." R.S. Berry 

and Hiro Makino, Technology Review, v. 76, February 
1974; pp. 32-43 606 

36. "Goods and services: an input-output analysis." David J. 

Wright, Energy Policy, v. 2, December 1974; pp. 307-315. 619 

37. "The energy costs of materials." P.F. Chapman, Energy 

Policy, v. 3, March 1975; pp. 47-57 . 628 

38. "An international comparison of polymers and their al- 

ternatives." R.S. Berry, T.V. Long, H. Makino, Energy 
Policy, v. 3, June 1975; pp. 144-155 639 

39. "The energy cost of goods and services." Clark W. Bullard III 

and Robert A. Herendeen, Energy Policy, v. 3, December 
1975; pp. 268-278 651 

40. "The energy cost of goods and services in the Federal 

Republic of Germany." Richard V. Denton, Energy 
Policy, v. 3, December 1975; pp. 279-284 662 



APPENDIX I 

Spreading Awareness 

From its origins in highly technical academic circles, energy ac- 
counting has caught the fancy of conservationists, scientists, and mak- 
ers of public policy. The articles which follow, starting with Howard 
T. Odum's seminal "Energy, Ecology, and Economics," typify the 
diffusion of awareness of the potential of this new technique into other 
disciplines and to the public at large. 



19 



Energy, Ecology, and Economics 



BY HOWARD T ODUM 



As long-predicted energy shortages appear, as questions 
about the interaction of energy and environment are raised 
in legislatures and parliaments, and as energy-related infla- 
tion dominates public concern, many are beginning to see that 
there is a unity of the single system of energy, ecology, and 
economics. The world's leadership, however, is mainly ad- 
vised by specialists who study only a part of the system at a 
time. 

Instead of a single system's understanding, we have ad- 
versary arguments dangerous to the welfare of nations and 
the role of man as the earth's information bearer and pro- 
grammatic custodian. Many economic models ignore the 
changing force of energy regarding effects of energy sources 
as an external constant; ecoactivists cause governments to 
waste energy in unnecessary technology; and the false gods 
of growth and medical ethics make famine, disease, and cata- 
lytic collapse more and more likely for much of the world. 
Some energy specialists consider the environment as an an- 
tagonist instead of a major energy ally in supporting the 
biosphere. 

Instead of the confusion that comes from the western civ- 
ilization's characteristic educational approach of isolating 
variables in tunnel-vision thinking, let us here seek common 
sense overview which comes from overall energetics. Very 
simple overall energy diagrams clarify issues quantitatively, 
indicating what is possible. The diagrams and symbols are 
explained further in a recent book (1). 

For example, TTgufe T shows' the basis of production in 
interaction of fuel reserves, steady energies of solar origin 
and feedback of work from the system's structure. Figure 1 
is the computer simulation of this model for our existence, 
showing a steady state after our current growing period. As 
the fuel tank is drained, we return to a lower solar base of 
simpler agriculture. Simple macroscopic minimodels based 



on overview of world energy provides the same kind of trend 
curves as the detailed models of Forrester and Meadows (see 
Ref 2). With major changes confronting us, let us consider 
here some of the main points that we must comprehend so 
we may be prepared for the future. 

1 . The true value of energy to society is the net energy, which 
is that after the energy costs of getting and concentrating that 
energy are subtracted. 

Many forms of energy are low grade because they have to 
be concentrated, transported, dug from deep in the earth 
or pumped from far at sea. Much energy has to be used di- 
rectly and indirectly to support the machinery, people, sup- 
ply systems, etc to deliver the energy. If it takes ten units 
of energy to bring ten units of energy to the point of use, then 
there is no net energy. Right now we dig further and further, 
deeper and deeper, and go for energies that are more and 
more dilute in the rocks. Sunlight is also a dilute energy that 
requires work to harness. 

We are still expanding our rate of consumption of gross 
energy, but since we are feeding a higher and higher per- 
centage back into the energy seeking process, we are decreas- 
ing our percentage of net energy production. Many of our 
proposed alternative energy sources take more energy feed- 
back than present processes. Figure 2 shows net energy 
emerging beyond the work and structural maintenance 
costs of energy processing. 

2. Worldwide inflation is driven in part by the increasing 
fraction of our fossil fuels that have to be used in getting more 
fossil and other fuels. 



20 



For explanation of symbols, see references and notes (5) 



Figure 1 A. Generalized world model of man and nature based on one- 
shot fossil fuel usages and steady solar work. Pathways are flows of 
energy from outside source (circle) through interactions (pointed blocks 
marked 'X' to show multiplier action) to final dispersion of dispersed 
heat. The tank symbol refers to storage. Here world fuel reserve storage 
helps build a storage of structure of man's buildings, information, pop- 
ulation, and culture. 




Heat Dispersal 



Figure 1 B. Graphs resulting from simulation of the model in Figure 1 A. 
Available world fuel reserve was taken as 5x10" kilocalories and 
energy converted from the solar input and converged into man's produc- 
tive system of growth and maintenance was 5x10' 6 kilocalories when 
structure was 10 1B kilocalories. Peak of structural growth was variable 
over a 50-year period depending on amounts diverted into waste path- 




Figure 1 C. 



The steady state observed in some simulations of Figure 1 A 
one as in the graph shown here. 




Figure 2. Energy flow diagram illustrating energy laws, and the difference 
between net and gross energy flows. 



© Work in 
Getting Energy and 
Concentrating 




If the money circulating is the same or increasing, and if 
the quality energy reaching society for its general work is less 
because so much energy has to go immediately into the 
energy-getting process, then the real work to society per unit 
money circulated is less. Money buys less real work of other 
types and thus money is worth less. Because the economy 
and total energy utilization are still expanding, we are misled 
to think the total value is expanding and we allow more 
money to circulate which makes the money-to-work ratio 
even larger. Figure 3 shows the circulation of money that con- 
stitutes the GNP in a counter-current to the energy flow. 

3. Many calculations of energy reserves which are supposed 
to offer years of supply are as gross energy rather than net 
energy and thus may be of much shorter duration than often 
stated. 

Suppose for every ten units of some quality of oil shale 
proposed as an energy source there were required nine units 
of energy to mine, process, concentrate, transport, and meet 
environmental requirements. Such a reserve would deliver 
1/10 as much net energy and last 1/10 as long as was calculat- 



221 



21 



ed. Leaders should demand of our estimators of energy re- 
serves that they make their energy calculations in units of 
net energy. The net reserves of fossil fuels are mainly un- 
known but they are much smaller than the gross reserves 
which have been the basis of public discussions and decisions 
that imply that growth can continue. 

4. Societies compete for economic survival by Lotka's prin- 
ciple (3), which says that systems win and dominate that 
maximize their useful total power from all sources and flex- 
ibly distribute this power toward needs affecting survival. 

The programs of forests, seas, cities, and countries survive 
that maximize their system's power for useful purposes. The 
first requirement is that opportunities to gain inflowing 
power be maximized, and the second requirement is that 
energy utilization be effective and not wasteful as compared 
to competitors or alternatives. For further discussion see 
Lotka (3) and Odum (1). 

5. During times when there are opportunities to expand one's 
power inflows, the survival premium by Lotka's principle is 
on rapid growth even though there may be waste. 

We observe dog-eat-dog growth competition every time a 
new vegetation colonizes a bare field where the immediate 
survival premium is first placed on rapid expansion to cover 
the available energy receiving surfaces. The early growth 
ecosystems put out weeds of poor structure and quality, 
which are wasteful in their energy-capturing efficiencies, 
but effective in getting growth even though the structures 
are not long lasting. Most recently, modern communities of 
man have experienced two hundred years of colonizing 
growth, expanding to new energy sources such as fossil fuels, 
new agricultural lands, and other special energy sources. 
Western culture, and more recently, Eastern and Third 
World cultures, are locked into a mode of belief in growth 
as necessary to survival. "Grow or perish" is what Lotka's 
principle requires, but only during periods when there are 
energy sources that are not yet tapped. Figure 3 shows the 
structure that must be built in order to be competitive in pro- 
cessing energy. 

6. During times when energy flows have been tapped and 
there are no new sources, Lotka's principle requires that 
those systems win that do not attempt fruitless growth but 
instead use all available energies in long-staying, high diver- 
sity, steady state works. 



Whenever an ecosystem reaches its steady state after peri- 
ods of succession, the rapid net growth specialists are re- 
placed by a new team of higher diversity, higher quality, 
longer living, better controlled, and stable components. Col- 
lectively, through division of labor and specialization, the 
climax team gets more energy out of the steady flow of avail- 
able source energy than those specialized in fast growth 
could. 

Our system of man and nature will soon be shifting from 
rapid growth as the criterion of economic survival to steady 
state non-growth as the criterion of maximizing one's work 
for economic survival (Figure 1). The timing depends only 
on the reality of one or two possibly high-yielding nuclear 
energy processes (fusion and breeder reactions) which may- 
or may not be very yielding. 

Ecologists are familiar with both growth states and steady 
state, and observe both in natural systems in their work 
routinely, but economists were all trained in their subject 
during rapid growth and most don't even know there is such 
a thing as steady state. Most economic advisors have never 
seen a steady state even though most of man's million year 
history was close to steady state. Only the last two centuries 
have seen a burst of temporary growth because of temporary 
use of special energy supplies that accumulated over long 
periods of geologic time. 

7. High quality of life for humans and equitable economic 
distribution are more closely approximated in steady state 
than in growth periods. 

During growth, emphasis is on competition, and large dif- 
ferences in economic and energetic welfare develop; com- 
petitive exclusion, instability, poverty, and unequal wealth 
are characteristic. During steady state, competition is con- 
trolled and eliminated, being replaced with regulatory sys- 
tems, high division and diversity of labor, uniform energy 
distributions, little change, and growth onl> for replacement 
purposes. Love of stable system quality replaces love of net 
gain. Religious ethics adopt something closer to that of those 
primitive peoples that were formerly dominant in zones of 
the world with cultures based on the steady energy flows 
from the sun. Socialistic ideals about distribution are more 
consistent with steady state than growth. 

8. The successfully competing economy must use its net out- 
put of richer quality energy flows to subsidize the poorer 
quality energy flow so that the total power is maximized. 



222 



AMBIO. VOL 2 NO. 6 



22 



In ecosystems, diversity of species develop that allow more 
of the energies to be tapped. Many of the species that are 
specialists in getting lesser and residual energies receive 
subsidies from the richer components. For example, the sun 
leaves on top of trees transport fuels that help the shaded 
leaves so they can get some additional energy from the last 
rays of dim light reaching the forest floor. The system that 
uses its excess energies in getting a little more energy, even 
from sources that would not be net yielding alone, devel- 
ops more total work and more resources for total survival. 
In similar ways, we now use our rich fossil fuels to keep all 
kinds of goods and services of our economy cheap so that 
the marginal kinds of energies may receive the subsidy bene- 
fit that makes them yielders, whereas they would not be 
able to generate much without the subsidy. Figure 4 shows 
the role of diversity in tapping auxiliary energies and main- 
taining flexibility to changing sources. 

Figure 4. Relationship of general structural maintenance to diversity and 
secondary energy sources. 




9. Energy sources which are now marginal, being supported 
by hidden subsidies based on fossil fuel, become less econom- 
ic when the hidden subsidy is removed. 

A corollary of the previous principle of using rich energies 
to subsidize marginal ones is that the marginal energy sour- 
ces will not be as net yielding later, since there will be no 
subsidy. This truth is often stated backwards in economists' 
concepts because there is inadequate recognition of external 
changes in energy quality. Often they propose that marginal 
energy sources will be economic later when the rich sources 
are gone. An energy source is not a source unless it is contri- 
buting yields, and ability of marginal sources to yield 
goes down as the other sources of subsidy become poorer. 
Figure 4 shows these relationships. 



10. Increasing energy efficiency with new technology is not 
an energy solution, since most technological innovations are 
really diversions of cheap energy into hidden subsidies in 
the form of fancy, energy-expensive structures. 

Most of our century of progress with increasing efficien- 
cies of engines has really been spent developing mechanisms 
to subsidize a process with a second energy source. Many 
calculations of efficiency omit these energy inputs. We build 
better engines by putting more energy into the complex 
factories for manufacturing the equipment. The percentage 
of energy yield in terms of all the energies incoming may be 
less not greater. Making energy net yielding is the only pro- 
cess not amenable to high energy-based technology. 

11. Even in urban areas more than half of the useful work 
on which our society is based comes from the natural flows 
of sun, wind, waters, waves, etc that act through the broad 
areas of seas and landscapes without money payment An 
economy, to compete and survive, must maximize its use 
of these energies, not destroying their enormous free sub- 
sidies. The necessity of environmental inputs is often not 
realized until they are displaced. 

When an area first grows, it may add some new energy 
sources in fuels and electric power, but when it gets to about 
50 percent of the area developed it begins to destroy and 
• diminish as much necessary life support work that was free 
and unnoticed as it adds. At this point, further growth may 
produce a poor ability in economic competition because the 
area now has higher energy drains. For example, areas that 
grow too dense with urban developments may pave over the 
areas that formerly accepted and reprocessed waste waters. 
As a consequence, special tertiary waste treatments become 
necessary and monetary and energy drains are diverted from 
useful works to works that were formerly supplied free. 

12. Environmental technology which duplicates the work 
available from the ecological sector is an economic handi- 
cap. 

As growth of urban areas has become concentrated, much 
of our energies and research and development work has been 
going into developing energy-costing technology to protect 
the environment from wastes, whereas most wastes are 
themselves rich energy sources for which there are, in most 
cases, ecosystems capable of using and recycling wastes as a 
partner of the city without drain on the scarce fossil fuels. 



223 



23 



Soils take up carbon monoxide, forests absorb nutrients, 
swamps accept and regulate floodwaters. If growth is so 
dense that environmental technology is required, then it is 
too dense to be economically vital for the combined system 
of man and nature there. The growth needs to be arrested 
or it will arrest itself with depressed, poorly competing econ- 
omy of man and of his environs. For example, there is rare- 
ly excuse for tertiary treatment because there is no excuse 
for such dense packing of growth that the natural buffer 
lands cannot be a good cheap recycling partner. Man as a 
partner of nature must use nature well and this does not 
mean crowd it out and pave it over; nor does it mean devel- 
oping industries that compete with nature for the waters 
and wastes that would be an energy contributor to the 
survival of both. 

13. Solar energy is very dilute and the inherent energy cost 
of concentrating solar energy into form for human use has 
already been maximized by forests and food producing 
plants. Without energy subsidy there is no yield from the sun 
possible beyond the familiar yields from forestry and agri- 



Advocates of major new energies available from the sun 
don't understand that the concentrations quality of solar 
energy is very low, being only 10 _lti kilocalories per cubic 
centimeter. Much of this has to be used up in upgrading to 
food quality. Plants build tiny microscopic semiconductor 
photon receptors that are the same in principle as the solar 
cells advocated at vastly greater expense by some solar 
advocates. The plants have already maximized use of sun- 
light, by which they support an ecosystem whose diverse 
work helps maximize this conversion as shown in Figure 5 A. 
If man and his work are substituted for much of the eco- 
system so that he and his farm animals do the recycling and 
management, higher yield results as in sacred cow agricul- 
ture (Figure 5 B). Higher yields require large fossil fuel 
subsidies in doing some of the work. For example, making 
the solar receiving structures (Figure 5 C), whereas the 
plants and ecosystem make their equipment out of the ener- 
gy budget they process. Since man has already learned how 
to subsidize agriculture and forestry with fossil fuels when 
he has them, solar technology becomes a duplication. The 
reason major solar technology has not and will not be a major 
contributor of substitute for fossil fuels is that it will not 
compete without energy subsidy from the fossil fuel econo- 
my. Some energy savings are possible in house heating on a 
minor scale. 



14. Energy is measured by calories, bni's, kilowatt hours, and 
other intraconvertible units, but energy has a scale of quali- 
ty which is not indicated by these measures. The ability to 
do work for man depends on the energy quality and quantity, 
and this is measureable by the amount of energy of a lower 
quality grade required to develop the higher grade. The 
scale of energy goes from dilute sunlight up to plant matter 
to coal, from coal to oil to electricity and up to the high qual- 
ity efforts of computer and human information processing. 



Figure 5. Diagrams of three systems of solar energy use. 




Figure S A. Man a minor part of the complex forest ecosystem. 



(b) Man's Diverse Work 
Substituting for 
Ecosystem Variety 




ri 



Figure 5 B. Man a major partner in an agricultural system on light alone. 




Figure 5 C. Fossil fuel subsidized 
technological society of man with 



a colonial member of 
possible solar conversion. 



AMBIO, VOL 2 NO. 



24 



15. Nuclear energy is now mainly subsidized with fossil fuels 
and barely yields net energy. 

High costs of mining, processing fuels, developing costly 
plants, storing wastes, operating complex safety systems, 
and operating goverment agencies make present nuclear 
energy one of the marginal sources which add some energy 
now, while they are subsidized by a rich economy. A self- 
contained, isolated nuclear energy does not now exist. Since 
the present nuclear energy is marginal while it uses the 
cream of rich fuels accumulated during times of rich fossil 
fuel excess, and because the present rich reserves of nucle- 
ar fuel will last no longer than fossil fuels, there may not 
be a major long-range effect of present nuclear technology 
on economic survival. High energy cost of nuclear construc- 
tion may be a factor accelerating the exhaustion of the richer 
fuels. Figure 4 illustrates the principle. 
Breeder Process: The Breeder Process is now being given 
its first tests of economic effectiveness and we don't yet know 
how net yielding it will be. The present nuclear plants 
are using up the rich fuels that could support the breeder 
reactors if these turn out to be net yielders over and be- 
yond the expected high energy costs in safety costs, occa- 
sional accidents, reprocessing plants, etc. Should we use the 
last of our rich fossil fuel wealth for the high research and 
development costs and high capital investments of pro- 
cesses too late to develop a net yield? 
Fusion: The big question is will fusion be a major net yield? 
The feasibility of pilot plants with the fusion process is un- 
known. There is no knowledge yet as to the net energy in 
fusion or the amounts of energy subsidy fusion may require. 
Because of this uncertainty, we cannot be sure about the 
otherwise sure-leveling and decline in total energy flows 
that may soon be the pattern for our world. 

16. Substantial energy storages are required for stability of 
an economy against fluctuations of economies, or of natural 
causes, and of military threats. 

The frantic rush to use the last of the rich oils and gas that 
are easy to harvest for a little more growth and tourism is 
not the way to maintain power stability or political and mili- 
tary security for the world community of nations as a whole. 
World stability requires a de-energizing of capabilities of 
vast war, and an evenly distributed power base for regular 
defense establishments, which need to be evenly balanced 
without great power gradients that encourage change of 
military boundaries. A two-year storage is required for sta- 
bility of a component. 



17. The total tendency for net favorable balance of payments 
of a country relative to others depends on the relative net 
energy of that country including its natural and fuel-based 
energies minus its wastes and nonproductive energy uses. 

Countries with their own rich energies can export goods 
and services with less requirement for money than those that 
have to use their money to buy their fuels. Those countries 
with inferior energy flows into useful work become subor- 
dinate energy dependents to other countries. A country that 
sells oil but does not use it within its boundaries to develop 
useful work is equally subordinate since a major flow of 
necessary high quality energy in the form of technical goods 
and services is external in this case. The country with the 
strongest position is the one with a combination of internal 
sources of rich energies and internal sources of developed 
structure and information based on the energy. The relations 
of energy sources to payment balances are given in Figure 6. 

18. During periods of expanding energy availabilities, many 
kinds of growth-priming activities may favor economic 
vitality and the economy's ability to compete. Institutions, 
customs, and economic policies aid by accelerating energy 
consumption in an autocatalytic way. 

Many pump priming properties of fast growing economies 
have been naturally selected and remain in procedures of 
government and culture. Urban concentrations, high use of 
cars, economic subsidy to growth, oil depletion allowances, 
subsidies to population growth, advertising, high-rise build- 
ing, etc are costly in energy for their operation and main- 
tenance, but favor economic vitality as long as their role as 
pump primers is successful in increasing the flow of energy 
over and beyond their special cost. Intensely concentrated 
densities of power use have been economic in the past be- 
cause their activities have accelerated the system's growth 
during a period when there were new energy sources to en- 
compass. 

19. During periods when expansion of energy sources is not 
possible, then the many high density and growth promoting 
policies and structures become an energy liability because 
their high energy cost is no longer accelerating energy yield. 

The pattern of urban concentration and the policies of 
economic growth simulation that were necessary and success- 
ful in energy growth competition periods are soon to shift. 
There will be a premium against the use of pump priming 



225 



25 



characteristics since there will be no more unpumped 
energy to prime. What did work before will no longer work 
and the opposite becomes the pattern that is economically 
successful. All this makes sense and is commonplace to those 
who study various kinds of ecosystems, but the economic 
advisors will be sorely pressed and lose some confidence 
until they leam about the steady state and its criteria for 
economic success. Countries with great costly invest- 
ments in concentrated economic activity, excessive trans- 
portation customs, and subsidies to industrial expansion will 
have severe stresses. Even now the countries - who have not 
gone so far in rapid successional growth are setting out to do 
so at the very time when their former more steady state cul- 
ture is about to begin to become a more favored economic 
state comparatively. 

20. Systems in nature are known that shift from fast growth 
to steady state gradually with programmatic substitution, 
but other instances are known in which the shift is marked 
by total crash and destruction of the growth system before 
the emergence of the succeeding steady state regime. 

Because energies and monies for research, development, 
and thinking are abundant only during growth and not during 
energy leveling or decline, there is a great danger that means 
for developing the steady state will not be ready when they 
are needed, which may be no more than 5 years away but 
probably more like 20 years. (If fusion energy is a large net 
energy yielder, there may be a later growth period when 
the intensity of human power development begins to affect 
and reduce the main life support systems of the oceans, at- 
mospheres, and general biosphere.) 

The humanitarian customs of the earth's countries now 
in regard to medical aid, famine, and epidemic are such that 
no country is allowed to develop major food and other crit- 
ical energy shortage because the others rush in their re- 
serves. This practice had insured that no country will starve 
in a major way until we all starve together when the reserv- 
es are no longer there. 

Chronic disease was evolved with man as his regulator, 
being normally as a device for infant mortality and merci- 
ful old age death. It provided on the average an impersonal 
and accurate energy testing of body vitalities, adjusting the 
survival rate to the energy resources. Even in the modern 
period of high encgy medical miracles, the energy for total 
medical care systems is a function of the total country's 
energies, and as energies per capita fall again so will the 
energy for medicine per capita, and the role of disease will 



Figure 6 A. Diagram showing how energy sources and energy loss path- 
ways affect the balance of payments and general economic competition 
position of a single country. Better balance results when one's own 
energy sources are better, and < 




again develop its larger role in the population regulation 
system. Chronic disease at its best was and is a very energy - 
inexpensive regulator. 

Epidemic disease is something else. Nature's systems 
normally use the principle of diversity to eliminate epidem- 
ics. Vice versa, epidemic disease is natures device to elim- 
inate monoculture, which may be inherently unstable. Man 
is presently allowed the special high yields of various mono- 
cultures including his own high density population, his paper 
source in pine trees, and his miracle rice only so long as he 
has special energies to protect these artificial ways and sub- 
stitute them for disease which would restore the high diver- 
sity system, ultimately the more stable flow of energy. 

The terrible possibility that is before us is that there will 
be the continued insistence on growth with our last energies 
by the economic advisors that don't understand, so that there 
are no reserves with which to make a change, to hold or- 
der, and to cushion a period when population;, must drop. 
Disease reduction of man and of his plant production sys- 
tems could be planetary and sudden if the ratio of popula- 
tion to food and medical systems is pushed to the maximum 
at a time of falling net energy. At some point the great 
gaunt towers of nuclear energy installations, oil drilling, and 
urban cluster will stand empty in the wind for lack of enough 
fuel technology to keep them running. A new cycle of dino- 
saurs will have passed its way. Man will survive as he repro- 
grams readily to that which the ecosystem needs of him so 
long as he does not forget who is serving who. What is done 
well for the ecosystem is good for man. However, the cul- 
tures that say only what is good for man is good for nature 
may pass and be forgotten like the rest. 



226 



iMBIO. VOL 2 NO 



26 



There was a famous theory in paleoecology called ortho- 
genesis which suggested that some of the great animals of 
the past were part of systems that vere locked into evolu- 
tionary mechanisms by which the larger ones took over from 
smaller ones. The mechanisms then became so fixed that they 
carried the size trend beyond the point of survival, where- 
upon the species went extinct. Perhaps this is the main ques- 
tion of ecology, economics, and energy. Has the human system 
frozen its direction into an orthogenetic path toward cultural 
crash, or is the great creative activity of the current energy- 
rich world already sensing the need for change? Are alter- 
natives already being tested by our youth so they will be 
ready for the gradual transition to a fine steady state that 
carries the best of our recent cultural evolution into 
new, more miniaturized, more dilute, and more delicate 
ways of man-nature? 

In looking ahead, the United States and some other coun- 
tries may be lucky to be forced by changing energy avail- 
abilities to examine themselves, level their growth, and 



change their culture towards the steady state early enough 
so as to be ready with some tested designs before the world 
as a whole is forced to this. A most fearful sight is the be- 
havior of Germany and Japan who have little native ener- 
gies and rush crazily into boom and bust economy on tem- 
porary and borrowed pipelines and tankers, throwing out 
what was stable and safe to become rich for a short period; 
monkey see, monkey do. Consider also Sweden that once 
before boomed and busted in its age of Baltic Ships while 
cutting its virgin timber. Later it was completely stable on 
water power and agriculture, but now after a few years of 
growth became like the rest, another bunch of engines on 
another set of oil flows, a culture that may not be long for 
this world. 

What is the general answer? Eject economic expansionism, 
stop growth, use available energies for cultural conversion 
to steady state, seek out the condition now that will come 
anyway, but by our service be our biosphere's handmaiden 
anew. 



H T Odum, Environment Power and Society (John Wiley) 336 pp. 
D H Meadows, D L Meadows, J Randry and W W Behrens III, The 
Limits to Growth (Universe Books, New York, 1972). 
A J Lotka, Contribution to the energetics of evolution in Proceedings 
of the National Academy of Sciences 8, 147-188 (1922). 
I am grateful for stimulation and collaboration of many in our common 
effort including especially C Kylstra, Pong Lem, and our keen graduate 
student group in the United States, and Jan Zeilon and Bengt-Owe 
Jansson in Sweden. Simulation work was supported by the U S Atomic 
Energy Commission on Contract Al-( 40-10-4398). 

Energy systems symbols used for showing mathematical and ener- 
getic relationships between the parts of our system of energy, enonom- 
ics and ecology. 



oO i 



Source Passive Heat 

Storage Sink 

All outside energy sources flow in from sources indicated with the 
circular symbol and these sources deliver causal forcing actions. All 
storages of energy, structure, money, information, value,, etc are rep- 
resented by the tank shaped symbol and these tanks are called state 
variables. All energies leave systems as dispersed heat that has no 
more potential for doing useful work. In the diagrams the dispersal of 
unusable heat energy is called a heat sink. 




Work Gate Self 



When two different kinds of flows of energy (or materials, information, 
or services that carry energy) interact in processes where both are 
necessary, we draw a work gale symbol. The system has an X if the 
action of one flow so facilitates the flow of the other and vice versa so 
that the process is a multiplier action. As in all processes, useful energy 
that drives the processes emerges as degraded, no longer reusable 



dispersed energy leaving the earth through the heal sink. (Heat on 
earth ultimately is reradialed out to space from the top of the atmo- 
sphere.) 

Self maintaining entities such as populations, cities, industries, and 
other organizations that feed energy from storage back into multi- 
plicative pumping actions are shown with the hexagonal symbol. The 
energy dispersed in maintaining the system, its growth, and its work 
shown passing out the bottom in a heat sink. 



T 



When new storages are developed, energy laws require that much of 
the energy be dispersed into unusable heat in order to make the 
process of storing go fast enough to be most competitive. The 
symbol lor potential generating work shows the necessary heal dis- 
persal that is required for any storing process. 

When two energy flows may be substituted for each other, we show 
their junction as the convergence of lines. This means that the flows 
add (in contrast to the work gate where other kinds of interactions 
are the result). 

Because money flows as a countercurrent to the flow of energy, goods, 
and services (the latter two also carrying energy), we represent path- 
ways that involve economic transactions with the diamond shape 
symbol and two counter diagrams pathways. The energy cost of doing 
economic business is shown as the energy lost into the heat sink. 
The diagrams may be examined as if they were a series of water tanks 
and pipes with water flowing between the tanks, being driven by the 
pressures of the storages or outside pressures and the energy of the 
water pressure ultimately leaving the system in the various frictional 
heat dispersions. The diagrams can thus be visualized to help see the 
complexity of systems and recognize just from the configurations what 
kinds of responses might follow proposed manipulatioas. As further 
given in ( 1 ) the diagrams are also ways of writing mathematical differ- 
ential equations for making precise mathematical descriptions of re- 
lationships. 



227 



[From Annals — American Academy of Political and Social Science, No. 410, 1973] 
An Energy Standard of Value 

(By Bruce M. Harmon) 

Abstract : The United States, as do most advanced industrial nations, generally 
measures value in money terms. The utility of employing a common denominator, 
such as money, is readily understood. However, within the past ten years there 
has been a growing disenchantment with money standards of measurement — 
particularly in the evaluation of public sector, nonmarket decisions. Concerns 
over distributive effects, regional consequences and environmental impacts have 
contributed to the belief among many that evaluative standards other than 
money ought to be adopted. Alternative proposals have been made for the estab- 
lishment and adoption of better measures for assessing developmental decisions. 
One alternative rests upon the assumption that energy is a critical variable in 
the post-industrial society of America ; energy costs in all areas of the produc- 
tive processes could thus be used in both public and private developmental de- 
cisions — operating and capital expenditure — to add another dimension to the 
traditional money standard of value. Furthermore, developmental matters could 
be judged not only in terms of dollar evaluations, but also in terms of BTUs the 
project would require. An energy flow model designed to detail the total energy 
cost of goods, and services for a given period in the United States — which has 
been developed at the University of Illinois — can serve to focus attention upon 
energy costs for various developmental undertakings. 

The American economy, as do those of other developed nations, generally meas- 
ures values in money terms. Gross national product (GNP) is widely under- 
stood as a monetary summary of national productivity ; until quite recently, 
cost-benefit analyses, particularly of public sector decisions sought to balance 
economic goods and bads in dollar terms. The utliity of a common denominator, 
such as dollars, needs no defense. In the last decade, however, there has been 
a growing disenchantment with dollar measures — particularly in the evaluation 
of public sector, nonmarket decisions. Concerns over distributive effects, regional 
consequences and environmental impacts have contributed to the recognition 
that GNP is not a sufficient measure. 

INTRODUCTION 

Various alternative proposals have been made for the development of better 
measures to assess developmental decisions. One approach has sought to identify 
social indicators as a statistical aid to program and policy choice. 1 Basically, 
the social indicator proponents argue that a double entry system of national 
accounting should replace the single entry, GNP calculus. More recently, in a 
related effort, some economists have attempted to develop methods for measuring 
net economic worth (NEW) . 

This article suggests another approach, one which rests on two assumptions; 
(1) that energy is a critical factor in the functioning of the American system 
and (2) that currently utilized energy supplies are finite and, in fact, are being 
consumed at an ever increasing rate. This approach suggests that public and 
private economic development decisions — both operating and capital expendi- 
ture decisions — should be assessed not only in dollar terms, but also in energy 
consumption terms — BTUs. Thus, where alternatives exist, choice in this sys- 
tem would rest on alternate energy requirements, as well as on alternate dollar 
costs and benefits. 



1 Social Goals and Indicators for American Society, The Annals 371 and 373 (May 1967 
and September 1967). 

(27) 



28 

At the University of Illinois Center for Advanced Computation (CAC), we 
have formed an energy research group — supported by the National Science and 
Ford Foundations — which is investigating the use of energy in the United States. 
The group is involved in three basic questions: (1) what is the energy cost of 
a good or service; (2) what are the alternatives to various goods or services 
which use less energy and ; (3) what is the dollar cost of, and what will the im- 
pact on employment and pollution be, if these alternatives are adopted. 

We really wish to determine how much energy could be saved throughout the 
entire production, delivery, operation and maintenance system if : as consumers, 
we switch to substitute products or selectively restrict our consumption ; as in- 
dustrialists, we switch to alternate processes ; and, as government policymakers, 
we regulate the flow of energy. Since each of these changes is reflected through 
the intricate web of the economic system, it is practically impossible for the in- 
dividual to perceive whether a given change increases or decreases overall energy 
consumption. Even if we could understand the net energy effects of possible 
changes, we would wish to rank them in increasing order of impact on our per- 
sonal and business lives. 

MODELING ENERGY USE 

During the energy research group's first year of existence an energy flow 
model — the CAC model — was developed. 2 It details the total energy cost of goods 
and services for 1963, the latest available data for the United States economy. 3 
The model provides the total, direct and indirect, use of energy by type — that is, 
gas, refined petroleum, coal and electricity — employment by occupation — 165 
types — and pollution — 10 types — for 362 sectors representing the industrial 
and commercial economy. In the CAC model direct energy is that consumed di- 
rectly by a particular industry to produce a unit of its goods and services ; 
indirect energy is that used by the suppliers of materials and services to the in- 
dustry and by the suppliers of these suppliers who supply only those materials 
which were needed for the unit of goods or services. Indirect energy is the limit 
of an infinite sum of terms which, although they increase in plurality, decrease 
in value. In some cases this process includes the amount of production of a 
particular industry used to make a unit of its own production — that is, a feed- 
back process which even includes the consumption of cars used by steel company 
executives to make the steel which is consumed in making a car. 

For a specified list of expenditures, the direct and indirect energy and em- 
ployment requirements and pollution generated in the industrial and commer- 
cial sectors can be determined for the technology used in 1963. We are currently 
developing data for 1967 to match the Department of Commerce dollar flow data, 
which will soon be available. At this point we can begin to understand how 
energy use changes with concommitant alterations in demands for goods and 
services. Thus, projections for future years become more feasible. 

AREAS FOR ENERGY CONSERVATION 

Many specific techniques to conserve energy can be imagined. One can discern 
three general categories from this plethora of opportunities : efficiency of pro- 
duction, efficiency of product use and control of the rate of energy use. In order 
to recognize the options for energy conservation, each of these can be thought 
of in the context of three broad classes of consumption : personal, government 
and industry. 

Production efficiency 

Because energy costs to the user are so low — only 3.6 percent of producers' 
price in 1963 — it is presumed by many that industries simply do not strive to 
use energy efficiently in their production processes. Compelling arguments for this 
point of view are made by Berg, 4 who claims that about 25 percent of the total 
United States energy use could be saved through more efficient use. For example, 
savings of up to 39 percent could be realized in the operation of certain equip- 
ment in the steel industry. 



2 R. A. Herendeen. "Use of Input-Output Analysis to Determine the Energy Cost of 
Goods and Services," document no. 69 (Urbana. 111. : Center for Advanced Computation, 
Universitv of Illinois. 4 March 1973). 

3 U.S., Department of Commerce. Input-Output Structure of the U.S. Economy: 1963 
(Washington. D.C. : Government Printing Office. 1969), vols. 1. 2 and 3. 

♦Charles Berg, "Energy Conservation thru Effective Utilization" (Washington. D.C: 
National Bureau of Standards, June 1972). 



29 

Railroads have improved their energy use efficiency by a factor of 10 since 
the early part of the century through both a change to diesel fuel and an im- 
provement of hauling techniques. 5 Recycling of aluminum, steel, paper, card- 
board and plastic offer rich energy saving opportunities. 6 However, the most 
ubiquitous energy increase in industrial processes is believed to have occurred 
via automation — that is, the displacement of labor from the production process. 
The ratio of production workers' wages to the cost of electricity increased steadily 
by 225 percent from 1951 to 1969. 7 During that time the wholesale price index for 
electrical machinery increased by 50 percent. 8 These factors indicate the pressure 
on the industrial decision makers to eliminate the increasingly expensive worker 
from the process and to substitute machines, which increase the energy intensity 
of the process. 

The energy research group has examined the general process of automation 
in some detail with the CAC model. If it is specified that each industry requires 
a one dollar increase in delivery to final consumption, then figure 1 shows the 
direct and indirect energy use and employment arising through the economic 
system. While a large proportion of the industries are centrally clustered, there 
are clearly some very energy-intensive industries — for example, asphalt coatings, 
asphalt paving, cement, primary aluminum, building paper and chemicals — and 
some very labor-intensive industries — for example, hospitals, hotels and credit 
agencies. The pattern shown in figure 1 represents the energy and labor require- 
ments of an additional dollar delivered to final demand. It represents, for a con- 
sumer, the direct and indirect effect on energy and employment of the expenditure 
of one dollar in each industry. It does not include the multiplier effects of the 
expenditure and, therefore, is inappropriate for use in an impact analysis. 



s 




■nsph codt 


O o 




.pacing -nix 


-J CNJ 






< rf 




. i .-lit 


? b^ 






u. 






o 






»- 






W o 






cc m 






5"j 






d o 






o 




• build paper 


Q 




.mi so print 


<T> 






£ o 

°> 00 

£ d 




, mi s< rhen, 






a. 




. steel 


3 




struct .lay . "bru* 


So 




zinc .' a lu.„ roll 






■plastics 


Z C\J 

O o 


ref petrol , 


. fit!* 1 
1 ■ ■ 11m 


5 


chcm min ■ 


paper. • frBJJfSg, 




• pulp .bottles 


If 


mlsc nonfer 


1 


£T O 






III 






z 






ID 


•• 


■ " % *.{..•• / » popery, ■ cotton oil 
.yj.fu, > .' tond »iik» 
.A, :.>V" ',>), ' poultry 'butter 

.'. ^SL^AS-h^ *., MMfed box- mUk ' cott 

■textim, . :VV*:V ■■' " ' - p rli??i* * ,rolts 


*8 

Q. 




forest prod 


? 8 

£8 


colfee • 


; . ■■-.. ■ ■ ■ -.^kn { 


homes' 





tobacco . 



3.4 6.8 10.2 13.6 17.0 20.4 23.8 27.2 30.6 34 

Totai Employmfni- Jo s Pft. 1963 Dollars to Final Demand (x 100.000) 

Figure 1. — Total — Direct and Indirect — Energy vs. Employment Intensities for 

362 Sectors in 1963. 
Soirce : Energy-Employment Policy Model, CAC, February 1973. 



5 Interstate Commerce Commission. "Transportation Statistics in the U.S.," annual 
report (Washington. D.C. : Government Printing Office). 

"Bruce Hannon. "System Energy and Recycling: A Study of the Container Industry." 
American Society of Mechanical Engineers. 72-WA-ENER-3 (New York, 1972). 

7 Bureau of Labor Statistics, Employment and Earnings, U.S., 1907-70. Bulletin 1312-7, 
table 5. Edison Electric Institute, Statistical Yearbook of the Electric Vtility Industry for 
1969 (New York. September 1970), p. 53. 

"U.S., Department of Commerce, Statistical Abstract of the U.S., 1971, 92nd ed. (Wash- 
ington, D.C. : Government Printing Office), p. 33(5. 



30 

Another way to consider the problem is to examine the effects of a 10 percent 
proportionate growth in each industry, with an offsetting decrease prorated 
among the other industries in proportion to their share of deliveries to final 
consumption ; thus, the GNP is conserved and the net multiplier effect of this 
differential type of growth is assumed to be nonexistent. In figures 2 and 3 first 
quadrant industries are primarily agricultural ; second quadrant industries are 
basic material production-, construction- and fabrication-oriented ; third quad- 
rant industries are service-oriented, with a high degree of technology and high 
wages ; and fourth quadrant industries are service-oriented, without a great 
degree of special labor saving technology and with low wages. Fifty percent of 
the industries fall in quadrant two, indicating that the nature of the structure of 
the 1963 economy was to respond to an increase in production by becoming more 
energy-intensive and less labor-intensive. Thus, figures 2 and 3 are addressed 
primarily to the policymaker concerned with the question of growth. The mag- 
nitudes reflect the relative dependence of United States society on each of its in- 
dustries in 1963. For example, a 10 percent increase in delivery to final demand 
by motor vehicles would have required a direct and indirect energy increase of 
34 trillion BTU and a decrease in employment of 104,000 jobs — direct and in- 
direct. Furthermore, a 10 percent increase in deliveries of postal services to final 
demand would have reduced energy consumption by about 4 trillion BTU and 
would have increased employment about 36,000 in 1963. Note that some inter- 
mediate products, such as steel and primary aluminum, deliver little to final 
demand. 



Pr role^re PeVintnf? 



CO 

CO o 

2 
UI 

DC 

O o 



< o 

o 



UJ *-" 

< 

I 

o 



« Chemicals 

* Gas Util. 

Motor Vehicles 

» Highway Constr. « Elec. Util. 

Water i . . 

m ,»iAir transport. 

Bldg. Constr. ? Trans.- 

Util. Constr.» ^^^r-.otor freight 

Misc. Chem. ' » » new CO nstr. xMghway "Water *c Sanitary 
* \V S yeggg gred jtgransp. * Non _ Fed Enterp. 

x « ^V"lluc Hotels 

Alcoh. Bevj. „",• * a » _ . office 

Pfo.Ser., *;««,»« BarbgrS Pers. Serv. xHospitals 

Auto Fep. w5iga?: Re -- profit Apparel (Pur.) 

Bai.ks x I .Insurance Carriers 
« Conanunicatipns 

« Doctors 



ifeeal Estate > Wholesale Trade 

Cvn Dwellings (-.U2,-19U) Retail Trade (.35,167) 



'-0.114 0.076 -0.038 0.000 0.038 0.076 0.114 0.152 0.190 

Change ir Employment, Million Jobs 

Figure 2. — Changes in Total Energy and Employment Requirements for a 10 
Percent Increase in Final Demand from the Noted Industry, Proportionately 
Absorbed from All Other Industries, 1963. 



31 

An inherent problem of this approach is the assumption that the gain in de- 
livery to final demand will be absorbed proportionately from all other indus- 
tries. Actually, the product of an industry competes with only a few other prod- 
ucts — for example, aluminum with steel and wood as structural materials or 
steel with glass and plastic as food containers. If one industry gained at the 
expense of a few competitors, the configuration of figures 2 and 3 would change. 
Suppose, for instance, that a 1 billion dollar gain in the primary aluminum de- 
liveries was obtained at the expense of an identical loss in steel deliveries. Then, 
from figure 1, energy use would increase about 116 trillion BTU, or by about 0.2 
percent of the total, and employment would decrease by 15,000 jobs, or by about 
0.03 percent of the total. An identical increase in primary aluminum deliveries at 
the proportional expense of all other industries would produce an increased use of 
energy of 332 trillion BTU and a loss of 65,000 jobs. 9 



3 

CD <o 
Z ™ 
O 

J 
-» 

CD 



a. 

5 

o 

u o 

cr d 

► 

ec 



a Toil. Prep. 
*F*rm Ma' h. 



» Plsis ;ce 
Railroad Cai & 



"Const. Mach. 
«2 



" San. Paper 



Sugar 



> Boilers 

* Syn. Ruhbe 

xCook. Oils 
•Convert Paper 
x Paper Mil In 



Canned Fruits ^ Veg. 



■Fabric Mills 



x House. 
»Feed Grain Refrig. 



flour x *» 

►Refrig. Macfi 



"Photo Kq. 
, Food Prep, n.e.c 






Confectionary _ . „ . , * "jiTv*rV^" 
' o_^..-*_ ' »-Soft Drinks * £ x ^ 1 



"Boat Bldg. 



•Forest FYoducts 



• Electric 

■(Telephones Instruments 



House Furn. , n.e.c. 
x Evap. milk 



Hosiery 
x iee :r eas 
i daii y farm prod. 



itJphol. Furn. 
> Leather Prod. 



A Wood fur niture 



x Tobacco 
Tobae. Dry. 



Jevelrj "Books 



Aircraft. F.qpt. ti.e.c 



Vef;. Farrr,s 



"Movi , 



5 1 



■3.4 -1.7 0.0 1.7 34 

Change in E.^plovmeni, Million Jobs 



6.6 



Figure 3.— Changes in Total Energy and Employment Requirements for a 10 
Percent Increase in Final Demand from the Noted Industry, Proportionately 
Absorbed from All Other Industries, 1963 * 
Source : Energy-Employment Policy Model. CAC. February 1973. 

Presently, the results of our research indicate that, in general, most United 
States industries are trading labor for energy— that is, becoming more energy- 

™JL?•x 1 ^ 0lk *o n ?TT B u HanI Vi^• "£ n Ener ^ v - Pollution and Employment Policy Model." docu- 
February 1973) "*' Center f or Advanced Computations, University of Illinois. 

•An enlargement of center portion of figure 2. 



32 

intensive and less labor-intensive. These industries, as well as their competitors, 
can be identified through the use of the CAC model. Thus, if economic growth 
is desired, it can be so guided as to minimize the impact on energy use and maxi- 
mize employment demands. In any event, the model clearly provides an estimate 
of the total energy and employment impact of desired shifts in demands. 

Several other process efficiency studies are either underway or have been 
completed : Moyers' study on the value of residential insulation ; 10 studies by 
Stein u and by Grot and Socolow u have shown the value of better heating and 
lighting techniques for both public buildings and private residences. However, 
these studies include only direct energy use. 

PRODUCT USE EFFICIENCY 

The variety of goods and services available in the United States today provides, 
according to Toffler, overchoice. 13 When the historical perspective of the develop- 
ment of the products is added, the variety of the system is even larger. For 
example, beer is currently available in twenty-three different package combina- 
tions ; at the same time, about five other configurations — including the consumer- 
owned container — have become obsolete. Another example of variety is the choice 
of intercity passenger transportation : plane, train, bus and car ; yet, the intercity 
passenger train is almost completely defunct. 

One can speculate that certain products are more energy efficient per unit of 
service than others. For instance, it has been shown that refillable bottles are 
about one-third less energy consumptive per unit of beverage than are paper, 
glass, aluminum or steel disposable containers. 14 Folk demonstrated that if the 
nation were to shift completely to returnable beverage containers, employment 
would rise by 130,000 jobs, and annual consumer costs would decrease by 1.4 
billion dollars. 15 The national energy savings would be about 0.5 percent, half of 
which would not be saved if the consumer savings were absorbed by an increase 
in average personal consumption. 

Several other product comparisons are now being investigated at the center. 
In particular, the intracity auto and bus are beiing compared. Results of the 
energy and labor cost studies of the auto in 1963 are presented in table 1. The 
auto consumed 12.4 percent of the GNP, required 12.0 percent of total employ- 
ment and consumed about 20.7 percent of total United States energy. This amounts 
to about 7.900 BTU per passenger mile and 5.5 jobs per 100,000 passenger miles. 
Preliminary estimates indicate that the bus is about one-third as energy-intensive 
as the auto in intracity operation, and that total United States energy use could 
be reduced about 5 percent by a full shift to buses in the cities, which is a signif- 
icant savings for a single change. This change is equivalent to a one-third increase 
in efficiency of operation of the current fleet of United States cars. An extremely 
interesting product alternative is the use of picture phones as a substitute for 
physical transportation. 16 Studies by Hirst highlight the energy conservation 
potential generally available in the transportation sector. 17 



10 John C. Moyers, "The Value of Thermal Insulation in Residential Construction : Eco- 
nomics and Conservation of Energy," 37830, Report ORNL NSF EP 9 (Oak Ridge, Tenn. : 
Oak Ridge National Laboratory, December 1971). 

11 Richard Stein, "Architecture and Energy" (New York : Stein and Associates, December 
1971). 

12 Richard Grot and R. H. Socolow, "Energy Utilization in a Residential Community" 
(Princeton, N.J. : Center for Environmental Studies, Princeton University. February 1973). 

13 Alan Toffler, Future Shock (New York : Random House, 1970), chap. 12. 

14 Hannon, "System Energy and Recycling." 

15 Hugh Folk, "Employment Effects of a Mandatory Deposit Regulation" (Chicago, 111. : 
Institute for Environmental Quality, January 1972). 

16 A. Goldsmith, "The Relationship of Telecommunications to Urban Transportation'' 
(Paper delivered at the Eighth Autumn Meeting of the National Academy of Engineering, 
12 October 1972). 

17 Eric Hirst. "Energy Consumption for Transportation in the U.S.." ORNL-NSF-EP-15 
'Oak Ridge, Tenn. : Oak Ridge National Laboratory, March 1972). 



33 

TABLE 1.— ENERGY AND EMPLOYMENT, DIRECT AND INDIRECT, FOR THE PRIVATE AUTOMOBILE IN 1963 



Category 



Final — 
demand, 


Energy 
Trillion 


Percent of 


Employment 




Thousands 


Percent of 


(billions) » 


Btu's 


total » 


of jobs 




total 


$5.86 


5,860 

1,220 

130 


57.7 
11.8 .. 
1.3 


278.8 




3.9 


4.05 


652.4 




9.1 


.83 


50 


.5 


39.5 




.5 


.55 


20 


.2 


88.6 




1.2 


14.43 


1,010 


9.8 


1, 155. 5 




16.1 


10.67 


350 


3.4 


1,718.7 




24.0 


.83 


80 


.8 


54.6 




.8 


.55 


20 
390 
580 


, .2 
3.8 .. 
5.6 


88.6 




1.2 


4.96 


471.6 




6.6 


8.96 


280 


2.7 


803.1 




11.2 



Fuel, produce ... 

Refining 

Retail 

Oil, produce 

Retail 

Automobile, produce 

Retail 

Tires, produce 

Retail 

Parking 

Highway construction (fuel taxes)... 
Insurance. 

Total »73.4 



» 10, 330 



100.0 



< 7, 149. 4 



100.0 



> Excludes household and government industries employment— 10,700,000 jobs. 

• 12.4 percent of total GN P. 

» 2C.7 percent of total U.S. energy used. 

♦ 12.0 percent of total employment 

Sources: Col. 2, R. A. Herendeen, "Use of Input-Output Analysis to Determine the Energy Cost of Goods and Services," 
document No. 69 (Urbana, III.: Center for Advanced Computation, University of Illinois, 4 Mar. 1973). Col. 3, Bruce, Hannon 
and S. Nakagama, 'The 1963 Direct Employment Intensity Vector," document No. 63, (Urbana: Center for Advanced Com- 
putation, University of Illinois, January 1973). 

Another interesting application of the concepts of product use efficiency, shown 
in table 2, is determining the various total energy and employment demands 
needed to supply a pound of protein. Cheese and fish are considerably more protein 
efficient — energy standard — than meat or milk in the forms consumed in the 
United States in 1963. Other product alternatives, such as food preparation and 
packing, clothing fibers and home appliances, are also being investigated for their 
unit energy and employment demands. The impact of such shifts on consumer 
cost, employment and pollution should be thoroughly understood before policy 
recommendations can be made. 

RATE OF ENERGY USE 

Ultimately, all the product and process energy efficiency gains may be inade- 
quate ; in this event the rate of energy use would have to be restricted. What are 
the priorities of restriction? How does the individual and the family draw upon 
the energy resource base? In which areas of use — direct and indirect — will an 
energy use restriction be least harmful? How does the direct and indirect 
energy use per family vary with socio-economic variables, such as age and income 
of the family head? 

TABLE 2.— TOTAL ENERGY AND EMPLOYMENT REQUIRED TO DELIVER A POUND OF PROTEIN TO THE CONSUMER 
THROUGH VARIOUS FOOD PRODUCTS IN 1963 



Food product 



Total employ- 
ment (16| 
Producer's Total production Production demands, (jobs 

price 1963 energy [6] to energy to per million 

(dollars per protein ratio energy content pounds of 

pounds) (koal/pound) ratio » protein) 



Meat products 

Cheese, natural and processed. 
Fluid milk 

Fresh or frozen packaged fish.. 



0.50 
.30 
.12 
.35 



32,600 
18, 800 
51,200 
17, 700 



6.3 
2.6 
6.1 
6.5 



» Marketing energy not included— add about 10 to 15 percent; total claoric energy used. Note that higher protein foods 
are not particularly calorie-rich. 

Source: Col. 1, U.S., Department of Commerce, "Statistical Abstr-» of the United States," 1971, approximate. 



34 

A study to determine the energy cost of different lifestyles is underway at the 
center. The first goal is to establish, through the use of equation (1), the influence 
of age, income and family size on the total energy budget of a family ; such 
studies will reveal the family dependence on total energy for housing, food, cloth- 
ing and transportation. If a priority of need is established, energy savings which 
would result from rationing could be estimated. Furthermore, the impact of an 
increase in energy cost on low income families could be estimated. Finally, studies 
such as these could be used to plumb the energy demands of varying degrees of 
affluence, leisure, convenience and variety of consumer products. 

A preliminary total energy and employment comparison of the average urban 
family — of about three persons — in the United States in 1950 and 1960 is made 
in table 3. Several interesting, tentative facts emerge : (1) energy use and employ- 
ment change are not proportional to change in constant dollar expenditures ; (2) 
direct energy consumption is about one half of the total energy demand; (3) 
each family generates about the expected direct and indirect employment, 1.1 per 
family, that the family itself provides — that is, working head of household and 
one out of ten spouses. 

TABLE 3.-PREU MINARY TOTAL (D1 RECT AND I NDIRECT) ENERGY AND EMPLOYMENT GENERATED BY THE AVERAGE 
UNITED STATES URBAN FAMILY IN 1950: 3.0 PEOPLE, AND IN 1960: 3.1 PEOPLE 









1950 






1960 








employment 






employment 






1950 energy 


demand 




1960 energy 


demand 




1950 dollar 


demand 


thousandth 


1960 dollar 


demand 


thousandth 


Item 


expenditure 1 


(million Btu ») 


of a job » 


expenditure^ 


(million Btu 3) 


of a job 2 


Food and beverage 


1,275 


77.0 


176.2 


1,288 


76.5 


173.7 


Housing' 


1,132 


262.1 


57.8 


1,536 


248.8 


74.0 


Clothing and personal 














care 


519 


30.4 


75.0 


592 


34.3 


86.5 


Medical care 


287 


11.1 


94.5 


362 


13.9 


118.5 


Recreation and education. 


287 


9.4 


39.5 


305 


10.6 


42.4 


Automobile and other 














transportation 


587 


73.6 


61.7 


779 


100.6 


80.4 


Contributions, insurance 














and other expenses... _ 


565 


19.4 


65.8 


1,071 


27.0 


88.9 


Savings 


99 






+158 


8.9 


17.7 


Taxes 


510 


9.3 


28.6 


810 


14.9 


45.5 


Retail margin 


475 


15.2 


75.0 


500 


16.4 


80.5 


Total 


5,538 


* 507. 5 


674.2 


7,090 


• 551.9 


808.5 


Corrected total e 




527.4 


1,080.0 




566.7 


1, 140. 









1 In 1963 dollars, approximate wholesale and retail margins removed to separate column. 

* ERG=CAC energy-employment policy model (February 1973). 
1 Direct energy use if weather-dependent. 

* 55.2 percent direct. 
« 50.6 percent direct. 

8 Corrected for difference in energy and labor productivity ratio, for household and government industry employment 
and for unemployment. 

Source: Data taken from Bureau of Labor Statistics, "Handbook of Labor Statistics, 1969" (Washington, D.C 1970,).' 
pp. 333, 400. 

The effect of family — of four persons — income level on energy demand is 
shown in table 4. Because of the preliminary nature of the estimate, the same 
energy coefficients were used for the same category in each income class; this 
procedure yields a total family energy demand which is roughly proportional 
to income. A more detailed estimate, now underway at the Center for Advanced 
Computation, will demonstrate how specific buying habits vary in each different 
Income class. The data of table 4 do reveal : energy dependency on food declines 
with income; energy dependency on housing rises with income; and energy 
dependency on transportation peaks with the intermediate income level. These 
three categories demand about 75 percent of the total family direct and indirect 
energy at each income level. The same three categories require 58 percent of 
the low budget income and about 47 percent of the other, higher dollar budgets. 
From this information we can infer that a rise in the price of energy would be 
preferentially difficult for the low budget family. 



35 



TABLE 4.— PRELIMINARY DIRECT AND INDIRECT ENERGY BUDGETS OF 3 DIFFERENT INCOME CLASS 4-PERIOD 

FAMILIES, 1970 





Low budget 


Intermediate budget 


High budget 








Energy, 




Energy, 




Energy, 






million 




million 




million 




Dollars 


RTV's 


Dollars 


Btu's 


Dollars 


Btu's 




(percent) 


(percent) 


(percent) 


(percent) 


(percent) 


(percent) 


Total 


6,960 


455 


10,664 


709 


15,511 


1,015 


Food 


22.5 


18.5 


18.8 


14.7 


16.3 


13.1 


Housing* 


19.1 


43.0 


22.3 


48.2 


23.1 


50.7 


Transport 


16.3 


10.8 


7.6 


12.5 


6.8 


11.3 


Clothing and personal 














care 


8.6 


6.8 


8.1 


6.2 


8.1 


6.3 


Medical care 


7.8 


4.1 


5.2 


2.6 


3.7 


1.9 


Other con umption 


4.0 


3.0 


4.9 


3.6 


5.6 


4.1 


Other costs 


4.6 
6.4 
10.0 


1.6 
5.2 
2.5 


4.8 
3.5 
14.3 


1.6 
2.7 
3.4 


5.6 
2.4 

18.4 


1.9 


Social security 


1.9 


Personal income tax 


4.4 


Retail (approximate) 


10.7 


4.7 


10.5 


4.4 


10.3 


4.4 



1 Proportioned from intermediate income to high and low budget using data from table 3. Energy will vary with climate. 

Sources: Cols. 1, 3, and 5, U.S. Department of Labor, "Handbook of Labor Statistics," Bulletin 1705, tables 126, 127 and 
128 (Washington, D.C., 1971). Cols. 2, 4, and 6, Center for Advanced Computation, Energy-Employment Model (Urbana, 
III.: University of Illinois, February 1973). Corrected for 1963 to 1970 energy productivity change. 

Finally, from the Bureau of Labor Statistics data 18 one can learn that the cost 
of living indices are higher than income indices for two-person families who live 
in cities, as opposed to the nonmetropolitan areas, and for those who live in north- 
ern, rather than in southern, United States cities. Intermediate values are found 
for north-central and western cities. Therefore, one can tentatively speculate 
that cooler climates and higher density living patterns stimulate the need for 
higher per capita total energy use. 

ACHIEVING ENERGY CONSERVATION 

The overriding counterforce to conserving energy is the apparently basic human 
desire for convenience. As one traces the history of United States industrial and 
personal activities, one eventually realizes that the trends indicate a drive for 
an easier or more convenient way of life. Such desires seem to have produced 
automation, higher physical mobility and diversity of products and services. In 
the face of such desires, the question then becomes : if we are truly entering a 
period during which energy supply cannot meet demand — an energy crisis — how 
do we reduce demand? 

Basically, there are two ways : (1) education, or socialization, programs which 
present the energy cost of alternative goods and services to the consumer and' 
(2) governmental regulation — through such measures as rationing and economic 
incentives schemes — of energy use. in addition to control of practices which lead 
to higher energy consumption. Education of the consumer can occur in many 
ways. The flow of information to consumers needs to be examined and, probably, 
to be modified. Education is supposed to occur in schools, but students probably 
learn more about consumption from the physical environment at the school than 
from classroom lessons. Interpersonal status gradients, insofar as they can be 
overcome by conspicuous consumption, exemplify part of the school environment. 
The mechanics of operating a school also must have an impact on the young 
consumer. For example, it is inconsistent — if not self-defeating — to teach general 
ecology in the classroom and to offer the students soft drinks in aluminum cans 
at lunch. 

Institutions, such as the school system and the government, can directly effect 
energy consumption not only by carefully reevaluating their physical plant proce- 
dures, but also, as major buyers of goods and services, by favorably influencing 
local and national market conditions toward lower energy -consuming practices. 
For example, the purchase of goods for the armed forces can be changed : cloth- 
ing variety can be reduced : less food packaging can be achieved through an 



18 U.S.. Department of Labor. Handbook of Labor Statistics (Washington. P.O. : Gov- 
ernment Printing Office. 1971). p. 302. table 138. 



36 

increase in the use of fresh foods, larger containers and less paper. The major 
institutions can have a signicant demonstration effect on the public at large. 

Media advertising — another form of public education — is obviously directed at 
increasing consumption. Product diversity is clearly the basis of advertising ; the 
energy inefficiency of diversity is increased by advertisers who convince the con- 
sumer to buy more or to buy a more expensive product or to do both. Yet, little 
is apparently known of advertising's effectiveness ; however, no major research 
effort is needed to determine the energy cost of the current information flow and 
the effects of controlling it to conserve energy. 

Regulation of direct and indirect energy can take five basic forms : energy ra- 
tioning, energy taxation and price control, information flow control, public invest- 
ment and land use control. Clearly, the federal government has the ability to 
ration fuels — as it did during World War II — and to reduce energy consumption 
by raising the cost, by adding taxes and/or by assuming control of the basic 
energy resource price. 

Regulation of the information flow — for example, control of advertising and 
product diversity — is perhaps the most benign process through which energy use 
can be controlled. If it is not sufficiently successful, other forms of more direct 
regulation — for example, gasoline rationing — would be required. Perhaps, direct 
regulation will be less oppressive and more widely understood if it follows an 
education program. 

For example, public investment in highways has probably caused a significant 
increase in automobile and truck traffic. Passengers and freight were formerly 
moved by the more energy efficient railroads — whose share of the intercity freight 
market has dropped from about 60 percent to 40 percent of the total in the last 
twenty years. 19 Although total freight hauled by railroads has increased in this 
same period, apparently, it has not been sufficient to provide enough revenue for 
adequate maintenance of equipment and roadbed. As another example, public 
electric utilities are so regulated that promotion of the use of electricity is re- 
quired : since the price of electricity is derived from the value of a utility's invest- 
ment, these companies opt for capital-intensive operations ; since the demand 
for eletcricity varies widely on a daily and seasonal basis, these large capital 
investments are regularly idled, causing the utilities to promote off peak use of 
electricity. Now seasonal peak use of electricity exists where minimum use once 
occurred. The phenomenon of air-conditioning was promoted by utilities in the 
fifties and early sixties to smooth out the demand curve and to provide efficient 
use of capital equipment. The former summer minimum demands *° are now the 
peak demands, and utilities advertise electric space heating — which is only half 
as efficient as direct gas heating — in an effort to fill the winter lows. Night time 
street and highway lighting is promoted to fill in the daily minimum periods of 
demand. 

As a final example, public investment policies might be used to control develop- 
ment of cities — thereby controlling total economic growth, or at least producing 
less energy-intensive land use patterns. A community could be planned around 
total energy systems in which waste heat from small power plants would be 
used to heat and cool the community's buildings. Work and home could be so 
located as to minimize transportation energy requirements. Intelligent land use 
planning is probably the most fundamental, long term key to energy conservation. 
Even elementary land use planning is in a juvenile state of development in the 
United States ; however, if we insist on increasing population and affluence, the 
only alternative to land use planning and regulation is the ubiquitous chaos 
present even now in and around our cities. 

AN ENERGY STANDARD OF VALUE 

Conventional energy — except for water power — is one of the nonrecyclable re- 
sources. Clearly, it is available in a finite supply, and the adverse environmental 
effects of its consumption are potentially great. The adoption of a national — and 
consequently a personal — energy budget appears to be necessary. The annual 
budget would represent a portion — dictated by our value of the future — of the 
proven energy reserves. Individual allocation could be similar to that of our 
present economies, which reflect personal value, except that we would have to 



19 Hirst. "Enerjrv Consumption for Transportation." p. 6. 

20 Hubbert Riss'er. "Power and the Environment — A Potential Crisis in Energy Sup- 
ply" (Illinois State Geological Survey, December 1970). p. 4. 



37 

strive for the right to consume energy ; the accrued currency would reflect the 
degree of success. The flow of the currency would be regulated by the amount 
of energy budgeted for a given period. If less energy existed at the end of the 
period, then currency flow would have to be reduced proportionately during the 
next period ; of course, an increase of currency flow would follow an abundance 
of energy. Recognition of the value of energy is equivalent to setting energy as 
the basis or standard of value. In doing so, society readmits itself into the natural 
system in which acknowledgment of energy's importance has never been lost. 



It Takes Energy to Produce Energy : The Net's the Thing 
(By Edward Flattau and Jeff Stansbury) 

Suppose you've found a wondrous goose that lays seven golden eggs a week. 
Does this mean you will be able to corner the gold market? Unfortunately, no, 
because to keep the bird fat and fluffy — and to keep its production up — you must 
feed it six golden egg yolks each week. Net result : only one golden egg for sale. 

This yarn may augur a potentially grim tale for the U.S. economy, for it sym- 
bolizes our net energy crisis. Not the various rigged shortages and price machina- 
tions you've been reading about, but the real thing : our net energy crisis. 

If you haven't pondered net energy, the fault's not yours alone. Senator Henry 
M. Jackson, Capitol Hill's leading energy warrior, hasn't heard of it either and 
William E. Simon, chief of the Federal Energy Office, wouldn't recognize a net 
energy ratio if he tripped over one on his way to a press briefing. Nor would The 
New York Times and The Washington Post reporters who have been following 
Simon around with almost a religious zeal. President Nixon, a flock of oil com- 
pany executives, and most influential economists have yet to discover net energy 
let alone apply its implacable logic to their decisions. 

Net energy is the energy you start with minus the energy you use up 
producing it — in other words, the calories you must spend to find, mine, trans- 
port, refine, convert, and deliver it. You must also add in any other energy yields 
that you have sacrificed in the singleminded pursuit of this one fuel. 

Without doubt, net energy may well be the simplest idea ever to have been 
ignored by so many acknowledged experts. Corporations would go bankrupt if 
they did not understand the distinction between gross income and net profit, 
but our thinking about energy has somehow not yet reached this level of sophis- 
tication. 

It does not matter whether you start with energy in the ground (coal, gas, oil, 
uranium oxide, plutonium, steam), in surface waters (hydropower reservoirs, 
tides, waves), and the land (timber, food, manure) or in space (sunlight, wind) : 
in order to use it, you must first expend almost as much energy just to obtain 
it. 

Agriculture presents an object lesson in the dynamics of net energy. Despite 
America's moderately high per-acre crop yields (up 63 percent in the last 20 
years), American farmers use more petroleum than any other economic group, 
and, as a result, they consume much more energy than they produce. 

California State University professor Michael Perelan has calculated that 
the energy value of the food Americans consume roughly equals the energy burned 
by tractors — just one fraction of our farm machinery. Of course you can't eat 
gasoline, so the equation is somewhat invalid. But what is significant is that 
the energy efficiency of U.S. agriculture has been steadily declining. Even ten 
vears ago some 150 gallons of gasoline per American flowed into food production 
That was five times as much energy as each of us consumed at the table — and the 
ratio has grown even worse in the ensuing decade. 

Perelman's statistics also show that most other countries balance this equation 
better than we do. In the extremely labor-intensive agriculture of China, a wet- 
rice fnrmer produces 50 units of food energy for each energy unit he expends. 
For a tvpical American grani farmers, the ratio is dramatically reversed : one 
unit of energy harvested for five expended. 

The largest portion of our energv deficit in agriculture comes from the use of 
nitrogen fertilizer. Since World War IT our per-acre yields of corn have tripled, 
but our nitrogen energy inputs have risen 10 times. That may be efficient in terms 
of man-hours, but it is wasteful of enenry. Dr. Georero Borgstrom. a food scientist 
Rt Michigan State University, calculates that it takes the calorie equivalents of 



38 

live tons of coal to make one ton of nitrogen fertilizer. Agriculture Department 
officials have stated that, because of the energy squeeze, we will face a nitrogen 
fertilizer shortfall of about one million tons this spring. And these shortages 
have already raised fertilizer prices by 30 to 60 percent over last year's level. 

Meanwhile, the sewage technology employed by our narrowly trained sanitary 
engineers has been dumping about 2.4 million tons of perfectly good nitrogen 
into our lakes, streams, and estuaries each year. Dr. John R. Sheaffer, until 
recently the U.S. Army's top environmental consultant, estimates that this nitro- 
gen is worth $1 billion and equals nearly a third of the synthetic fertilizer sold. 
By wasting it and forcing us to replace it with manufactured nitrogen, sanitary 
engineers make us burn the equivalent of an extra 2.2 billion gallons of fuel oil 
each year. Reclaiming most of this sewage will greatly improve the net energy 
yield of our agriculture, although it is not the whole answer. 

Howard T. Odum, professor of Environmental Engineering Sciences at the 
University of Florida, has done some of the most provocative thinking on the 
importance of net energy. "Many forms of energy are low-grade because they 
have to be concentrated, transported, dug from deep in the earth or pumped from 
far at sea," Odum says. "If it takes 10 units of energy to bring 10 units of energy 
to the point of use, then there is no net energy. Right now we dig further and 
further, deeper and deeper, and we go for energies that are more and more dilute. 
We are still expanding our rate of consumption of gross energy, but since we are 
feeding a higher and higher percentage back into the energy-seeking process, 
we are decreasing our percentage of net energy production." 

That single paragraph makes more sense about our energy predicament than 
volumes of solemn declarations from the Senate Interior Committee, the Federal 
Energy Office, and the oil companies. It puts tbe spotlight on the steady decline 
in the net energy yields of our traditional fossil fuels. Our nearest, least 
resistant oil fields have been drained ; now we must literally squeeze oil from 
marginal ones, or import it from abroad in tankers (which themselves use fuel), 
or pump it long distances from offshore rigs to refineries. In the past, when oil 
returned a high net energy yield, it handsomely subsidized all our other fuel 
and power sources. Oil built hydroelectric dams. Oil and electricity extracted our 
coal. Oil grew our wheat. Oil made 20th-century America "work." But today, our 
free ride on oil is coming to a halt. 

FLUNKING THE NET ENERGY TEST 

This means, among other things, that each of our remaining fuels must hence- 
forth meet the test of its own intrinsic net energy ratio. Let's look briefly at 
three such fuels : coal, uranium oxide, and shale oil. 

In a recent interview with The Washington Post, William Simon declared: 
"Today we've got an 800-year supply of coal in this country where we can get 
from 25 to 35 percent of our needs . . . for an infinite period of time." The 
kindest phrase we can think of to describe this statement is "whistling in the 
dark." Simon has simply commandeered — and inflated — the best prevailing esti- 
mates of our gross coal reserves without making any allowance for the energy 
cost of extracting this coal. 

Unfortunately, common sense suggests that the net yield from our remaining 
coal deposits will be low. Unlike oil from wells, coal can be extracted only after 
tons of earth have been pushed aside. In the case of strip-mined coal these tons 
are literally mountains — mountains which must later be bulldozed back into 
shape, covered with topsoil, seeded, and carefully tended for 15 or so years (that 
is, unless we're willing to leave the land unreclaimed) . 

Seven months ago the President's Council on Environmental Quality issued 
a booklet on electric power. It contained some serious errors, but one useful fact 
can be found in it. Of all the deep or strip-mined eoal that we extract, only SO 
percent reaches the final user. The rest disappears through physical losses in 
processing and transportation, heat losses during conversion to electricity, and 
electrical leakage from transmission lines. Furthermore, each of these steps, 
like the extraction process itself, consumes considerable energy. "Big Muskie," a 
giant coal extractor now tearing up Muskingum County, Ohio, gulps enough power 
each day for 27,000 all-electric homes. The energy costs of mining coal and 
converting it to electric power must then be subtracted from the 30 percent left 
after leakage and physical disappearance. 

We must also calculate the potential energy we lose when we decide to use 
land for strip-mining. Joseph Browder, of Washington's Environmental Policy 



39 

Center, says, "The energy costs of stripping Northern Great Plains coal must 
include the direct loss of agricultural productivity in the Powder River Basin 
and other areas where livestock produced on native grasses would be replaced 
by livestock produced in a feedlot system dependent on energy-intensive, fertilized, 
irrigated crops." 

In the West, the energy problems caused by strip-mining are a little different. 
Immense amounts of water will have to be set aside to gasify, liquefy, or convert 
its costs into electricity. This means new water sources will have to be developed, 
at a high energy cost, to replace those consumed by producing coal. 

When all these relevant energy costs are deducted, the net yield from strip- 
mining our deep coal reserves may drop to three or four percent or less. Simon's 
800-year bonanza has shrunk to a few decades. 

LOSING ENERGY WITH ATOMIC POWER 

Then there's nuclear power, which the Nixon Administration is betting will 
eliminate our energy problems forever by the end of the century So. far, how- 
ever, 25 years of nuclear fission power has drained off more energy than it has 
produced. It's easy to see where the problem comes from. 

In 1973, the Atomic Energy Commission (AEC) used 25.7 billion kilowatt-hours 
of electricity just to produce the uranium needed to fuel nuclear power stations — 
stations with a power output of about 50-billion kwh. This was not a fluke. "As 
much as half of the gross electrical output of a nuclear plant would have to be 
recycled to supply input for fuel processing," says E. J. Hoffman, a University 
of Wyoming nuclear energy specialist. If you also include the energy costs of 
searching for uranium ore, mining it and transporting it ; finding, mining, 
refining, and transporting the metals that go into nuclear power plants ; manu- 
facturing the concrete for these plants; operating them (which includes driving 
to and from work) ; and storing or reprocessing the "dirty" wastes from nuclear 
fission, you will find the 1972 net energy yield from uranium has probably 
sagged to less than 10 percent. 

The sober conclusion seems to be that our nuclear energy program would 
collapse without its big energy subsidy from oil. Three years ago, after studying 
the past, present, and projected yields of U.S. nuclear power plants, Dr. Hoffman 
concluded, "The cumulative energy expenditure of the entire atomic energy 
program may not be recouped from nuclear fission power plants by the time 
the reserves of economically recoverable U-235 are used up." 

Shale oil presents us with another statistical no-man's land. There are between 
one billion and 10 billion barrels of shale oil buried in the West, according to 
official estimates. However, after all the energy requirements of shale oil have 
been subtracted, there may be a net yield of only a few hundred million barrels. 
This could even drop off close to zero if the land from which the shale is extracted 
is to be pushed back into place an^ 1 recon toured. In a recent book. The Energy 
Crisis, Tawrence E. Rocks and Richard Runyon predict that the net energy yield 
from 99 percent of our Rocky Mountain oil shale will be zero. 

One reason why almost nobody seems to care about net energy is the tendency 
of otherwise knowledgeable people to confuse net energy with mere efficiency 
of extraction. When we hear the terminated efficiencv," it is easy to think of 
extraction rations : for example, that we only suck about 30 per cent of the oil 
in an oil field out of the ground, or that solar-energy devices can convert only 
about one per cent of the incoming light into useful energy. This may make oil 
seem more "efficient" than solnr energr. But it doesn't cost us anything: to pass 
up that 99 per cent of solar energy or 70 per cent of oil ; we get nothing and 
we expend nothing, so this loss doesn't really affect our calculations. What 
really counts is how much energy we must expend to get the obtainable energy. 
Extracting and refining the 30 per cent of the oil may consume so much energy 
that the net energy yield is quite low. 

Another problem is that the entire debate over fuels is dominated by eco- 
nomists, geologists, and capitalists, who have been trained to think only in 
terms of dollars. Is oil shale still too costly to develop? Well then, they advise, 
wait for conventional fuels to grow scarce (high-priced) and shale oil will he- 
come a profitable commodity. The trouble with such incantations is that they 
assume the dollar costs of a given fuel will faithfully — and quickly — reflect its 
energy costs. This assumption is unwarranted. The dollar costs of new oil wells, 
for example, depend on tax write-offs and other accounting decisions which do 
not reflect net energy. A fuel can remain artificially underpriced for months, even 



40 

years, after its net yield drops. And the notion that shale oil will become pro- 
fitable at the very moment when it can no longer count on a big initial subsidy 
from deep-well oil is, in net energy terms, absurd. 

"The truth is often stated backwards by economists," says Dr. Odum. "Often 
they propose that marginal energy sources will become economic when the rich 
sources are gone. But the ability of marginal sources to yield goes down as 
the sources of subsidy become poorer." 

Three months ago, Brookings Institution President and former Budget Direc- 
tor Kermit Gordon admitted to the American Economics Association. "I know 
of no neat theory of inflation that fits the facts of the last five years — neither 
aggregate demand, nor money supply, nor labor power, nor oligopoly power, nor 
bottlenecks, nor expectations — though I could easily be convinced that they all 
played a part." 

The "facts" of the last five years boil down to a lickety-split inflation that has 
respected neither boom nor bust. Without sounding pompous, it seems quite 
plausible that our shrinking net energies provide part of the theory that resolves 
this seeming paradox. Professor Odum sees it that way : "If the energy reach- 
ing society for its general work is less because so much energy must go into the 
energy-getting process, then the real work of society per unit of money circulated 
is less. Money then buys less real work. . .and is worth less." This inflationary 
bind could exist in a rising or a falling economy, provided that the actual money 
supply did not substantially dwindle. 

The infamous U.S. -Soviet wheat deal two years ago provides a vivid ex- 
ample of the fickle relationship between dollars and net energy. This ex- 
change was a financial windfall for a few grain exporting firms, but an energy 
disaster for Americans. Wilson Clark, a Washington, D.C. energy specialist, 
claims that it costs us 10 units of energy to ship grain worth one unit of energy 
to the Russians. In return we were promised Soviet natural gas worth two 
energy units. "Financially it worked out fine," Clark says, "but in energy terms 
we suffered a 5 to 1 net loss." 

BALANCING OUR ENERGY ACCOUNTS 

Clearly, before the United 'States invests hundreds of millions and then billions 
of dollars in a desperate scramble for miracle fuels, we must devise a system of 
national energy accounts. How much of the energy from western coal must be 
pumped back into its production? How will the net yield of western strip- 
mined coal compare with the midwestern deep-mined coal? What natural 
energy harvests will be sacrified by these alternatives? How heavily must we 
subsidize the next 25 years of nuclear fission? To bring one 1,200-calorie loaf 
of cracked wheat bread to a suburban table, how many thousands of calories 
must we spend on fertilizer, pesticides, cultivation, harvesting, farm overhead, 
milling, baking, distribution, sales, and — not least — that luxurious trip to the 
supermarket? For our overseas oil, how much energy do we invest in our 
Mediterranean fleet, in our farflung corporate empires, and in their support struc- 
tures in the Departments of State, Commerce, and Interior? Will solar con- 
verters yield a rich net energy harvest? And (assuming the necessary copper 
can be scavenged ) will windmills be practical ? 

A bookkeeping system capable of answering these and a vast number of 
related questions will do something for us that mere dollars cannot : it will 
test our economic sanity, rationalize our economic planning, and give us a long- 
lost sense of proportion vis a vis the natural world we inhabit. We need to 
launch this accounting revolution immediately, for the world we face tomorrow 
is not the world we know today. Immutable laws of net energy are leading us 
toward an economic steady state. 

"Our system of man and nature will soon be shifting from rapid errowth to 
steady state non-growth as the criterion of economic survival," says Dr. Odum. 
Ecologists are familiar with both the growth state and the steady state: they 
observe both in natural systems." Economists, however, have been schooled dur- 
ing, by, and for growth. Most of them have never seen a steady state. Except 
for the London School's Ezra Mishan and a mere handful of kindred spirits, 
they reject the possibility of a steady state even though man lived in something 
very close to one during 99 per cent of his evolution. 

Why is a new steady state — presumably at a decent level of health and well- 
being in our cards? It is in our cards because the high net-yielding energv 
sources we need to survive, with the doubtful exception of nuclear fusion, can- 



41 

not match the total daily output we have heretofore enjoyed from the fossil 
fuels. Oil, coal, and gas have been a marvelous energy "capital," a 400-million- 
year-old bankroll for the Western world. Sunlight is energy "income," however, 
we can tap only so much of it each day. Whether we like it or not, we'll have to 
live within our means. This is the only way we can reach that redoubtable state 
Mr. Nixon calls "energy self-sufficiency." 



Systems of Energy and the Energy of Systems 
(By Thomas A. Robertson) 

Burmah, the second largest oil company in Britain, falters, threatening the 
development of much-needed North Sea oil for that nation's troubled industries. 
Texaco, one of the largest U.S. oil companies, drops its plans to develop oil shale, 
a highly touted domestic energy alternative. Inflation punches up through 20 per- 
cent in Japan and Italy. Elsewhere, it continues to contradict predictions of 
lower rates. 

In Washington, Senator Mark Hatfield and Congressman Mike McCormick 
clash in a committee conference on energy research and development. The dis- 
cussion concerned an obsecure concept called "net energy." Few recognize the 
relationship of net energy to the problems of oil companies and economies. Even 
fewer appreciate the importance of this concept to the stability and well-being of 
the industrial society in which we live. Senator Hatfield insisted the concept 
of net energy should be one of the criteria in a selection of energy research and 
development. Congressman McCormick said he was concerned that the evaluation 
of energy technology for its net-energy potential would be restrictive to energy 
research. Mr. Hatfield, the Senator from Oregon, held his ground and the section 
on net energy was retained in the bill. This brief encounter was a milestone in 
the history of industrial society. In those few minutes, for the first time, two 
leaders touched on the keystone issue of how this society is powered. 

The term net energy first trickled into the consciousness of industrial society 
in the fall of 1973, when Howard T. Aodum, Graduate Research, Professor of En- 
vironmental Engineering at the University of Florida, was asked to submit a 
paper to AMBIO, the publication of the Royal Swedish Academy of Science. The 
result, "Energy, Ecology, and Economics," came out in December, 1973. As copies 
of the paper were circulated (it was reprinted in some 14 publications through- 
out this country and abroad), the concept of net energy made a strong impres- 
sion on concerned readers looking for better opportunities in these changing 
times. The concept of net energy is a product of Professor Odum's energy sys- 
tem analysis. Energy is used by Odum and his associates as a common denom- 
inator. Symbols of energy processes and of connecting flows track energy in its 
many forms to show the workings of all systems and combinations of systems. 
From this whole-system view of energy, environment, and economics, concepts 
such as net energy and a host of other significant, insights emerge. 

When Paul Samuelson, a Nobel Laureate in economics, says, in a recent Busi- 
ness Week article, "I think the greater error in [economic] forecasting is not 
realizing the other possibilities," he is making a strong case for questioning our 
existing processes of perception. Economists, using symbols called money to 
understand the allocation of resources and work, cannot see those things in 
the systems of society and nature that money does not immediately track. For 
example, money — the symbol — shows little about the services provided by the 
environment for which no money payments are made. "If you don't pay for it, 
it doesn't exist," is perhaps too harsh a characterization of the economic point 
of view, but it does make the point. 

The delayed effects through our economy caused by the rippling out and back 
of increased primary energy prices for oil, coal, and gas are dynamic properties 
of our social economic system about which little is known. When our leaders say 
a $3.00-per-barrel import tax will cost the consumer only $275 a year, they are 
isrnorinff how the butcher, the baker, and the candlestick maker (candle wax 
^omes from paraffin, a derivative of crude oil) pass their costs on to the consumer. 
The «o-ealled "free market." which economists and others sj)eak of. is so sur- 
rounded by confusion that it is hard to fiml any real meaning in the term. There is 
certainly little that is "free" about the flow of money resources and work through 



-3^1 O - 7(1 



42 

industrial society. Government with its immense regulatory ability, corporations, 
unions, and consumer advocates all seek to do what they think is best for them- 
selves. By the various means of legislation, lobbying, and associations, they 
disturb any "free" flow through the system. Truly beneficial results are increas- 
ingly difficult as the system in which they operate becomes more and more com- 
plex. Social good, corporate good, and even governmental good, to say nothing 
about good government, become more and more elusive. The counterproductive 
tendencies of a high-energy industrial system reveal today's profit as tomorrow's 
losses. 

Energy systems analysis lowers this perceptual barrier by integrating all parts 
of the systems under consideration first into conceptual approximations, and then 
into accurate simulation of the essence of the problem at hand. The result of this 
process of inquiry tends to be forced by the realities of the system and is much 
less dependent on the perceptual bias of the observer. For this reason, net energy 
appears to be one of these unforeseen systems circumstances. 

NET ENERGY AND INFLATION 

Net energy begins as a simple concept. It takes energy to get energy. What 
counts for use by society is the net energy "profit" from the work we (society) 
do to extract the given supply of energy. The accounting must be done in terms 
of both energy and money. Money alone as an accounting medium is not working. 
In other words, net energy is the amount of energy available from a given resource 
for use by society after subtracting the energy required to search for, extract, 
process, and transport the energy derived from that resource. 

The energy/dollar problem is one involving two different but not separate 
functions in our economy, and the best way to understand this subtle distinction 
is first to consider energy alone. Our effort as a society to find, process, and use 
fossile fuels can be likened to that of a family fueling its members with food. In 
our case, the family at first lives next door to a grocery store that is fully stocked 
but charges nothing for the food. The familys only cost is the energy they burn in 
walking to and from the store. As long as the store is nearby and the trip is short, 
the family is unaware of any significance "price" and happily assumes either (a) 
that the store will never be exhausted, or (b) that another full store will spring 
up alongside the first by the time the first runs out of food. 

Unhappily, the store runs out, and no new store takes its place. ( Several stores, 
which we might liken to alternative energy sources such as solar and nuclear 
power, appeared, but none had yet proved to contain any appreciable supply of 
energy /f ood. ) The family must now go to a store several blocks away — a trip that 
begins to exact a noticeable amount of food-energy cost to the family. Eventually, 
the only stores that can be found still stocked with food the family must have 
are a half-day's trip away. One day, the family realizes that it is spending the 
same amount of energy in traveling to and from the distant store as is contained 
in the food they pick up during the trin. There is, in other words, no net energy. 

Now add dollars to the above story. We start with each unit of energy havinsr an 
equivalent unit of money attached to it. The familv gets money for all work it 
does outside of going to the store. After the family get food/energy from its 
nearby store, it is able to use its surplus energy to do nonstore-going work for 
which the family receives money. With this money the family can buy still other 
kinds of work. (Work, in this sense, means the goods and services available from 
society.) The non-store-going energy is net energy. It is easy to see that as the 
family spends more and more of its energy groing to those more distant stores, it 
has less and less energy to do its money -making work. 

Several variations on this theme occur. In the example above, if we do le c s work 
we get less money. This is what would happen if we had that "free market" the 
economists talk about. However, you and I are part of a representative govern- 
ment that would find it hard to accept such a rigorous relationshin between energy, 
money, and the work we do. Governments, in order to make things look better, 
"grow" by adding money to that which is alreadv in the system. This works as 
long as there are resources available to back up the money. However, if resources 
are diminishing, this only makes us feel as if we have more ability to buy work, 
and only delays the time when we must reckon with reality. Understanding the 
basic elements; of net energy is like learning to ride a bicycle. Once you learn, you 
cannot believe how hard it was to get started, at the same time, you will never 
forget it. 



43 

The fundamental cause for inflation oan be seen as changes in net energy. As 
our concentration of resources diminish (the stores are more distant and harder 
to ge to), we do more work to bring in less and less net energy. Consequently, the 
amount of work done per unit of money diminishes. Three associated causes for 
inflation are : 

(1) Increased primary energy (crude oil, coal, and gas) prices. An example is 
the OPEC price hike in 1973 and the coal-strike settlement in December, 1974. 
Crude oil imports, as well as excise taxes, also fit this category. 

(2) Increasing the money supply without having the energy to do this addi- 
tional work. Examples are dropping the margin on the stock exchange, increases 
in credit such as reducing the reserves over which banks loan money, forcing 
interest down, and putting money into pump-priming projects without the ability 
to power them. 

(3) Decreasing productivity : in other words, getting less work from the energy 
we have. As economic writer Hazel Henderson says, "We are creating transac- 
tional costs faster than we are producing output." Everyone has his pet example 
of this. 

ENERGY-INVESTMENT RATIO AND LEVERAGE 

The net energy available to an economic system can be seen as an energy return 
on the energy invested. This "energy -investment ratio" changes over time as con- 
centrated resources become more dilute. A hard point to believe, but that we must 
consider, is that a time finally comes when your energy-investment ratio is so 
diminished you can no longer do the things you could do in the past. Businessmen 
speak of leverage, by which they mean investment ratio. Energy-investment ratio 
is the basic leverage that determines the winners and losers (if there ever are 
any) between competitors, be they individuals, corporations, or nations. No 
businessman would venture into any enterprise without being aware of the 
comparative leverage between himself and his competitors. 

Where money was working well in a stable economy, it moved energy in all 
its forms of resources, goods, and services, and it accurately and effectively 
allowed us to account for all the processes of industrial society in which we are 
involved. Modifications we now find in the availability of resources, particularly 
energy, are symptomized by inflation, which causes a deterioration in the quality 
of information we use in business and finance to move resources in our economic 
system. The energy-investment ratio and other elements of energy systems anal- 
ysis form a new economics. Using energy as an accounting medium along with 
money can re-establish the information quality of our economic system so neces- 
sary to the best understanding and use of scarce resources by our society. 

CHANGING ENERGY-INVESTMENT RATIO 

The best indication that a change is taking place in the United States energy- 
investment ratio has come from a paper done by M. King Hubbert of the U.S. 
Geological Service for Senator Henry Jackson's National Fuels and Energy Policy 
Study. In Hubbert's paper, which reviews fossil fuel energy availability for the 
future, we find a chapter titled "Discoveries per foot of Exploratory Drilling." 
Looking for an indication of work done for energy returned, as an energy- 
investment ratio, Hubbert found that, "The rate of discovery [of oil] is subject 
to wide fluctuations in response to extraneous conditions such as economic and 
political influences. In fact, the rate of discovery may be increased to maximum 
or shut down completely in response to managerial or political fiat, or to the 
changes in the economic climate." But in pursuing his investigations Hubbert 
found that the amount of oil discovered per unit of depth of exploratory drilling 
is almost exclusively a technological variable and is highly insensitive to economic 
or political influences. For example, while the officials of a large oil company can 
authorize its staff to double the amount of exploratory drilling in any given year 
and thereby increase the discoveries per year, no oil company management can 
successfully order its staff to double the quantity of ail to be found per foot of 
exploratory drilling. 

Hubbert's figures raise vital questions about our energy return on energy 
investment for oil production in this country. His reports suggests an 870-percent 
reduction in the return on our energy investment over the past few decades. Here 
is a fact that bears directly on the ability of the nation, or the whole of industrial 
society, to power itself. For all the discussion of national energy policy and the 



44 

current economic crisis, how many policy makers do you see touch on the essence 
and magnitude of the problem suggested here? 

Hubbert touches on the question of net energy almost by accident, but it is a 
sound start. The next step is for Odum and others to refine the applications of 
systems analysis to all the existing and proposed energy and economics circum- 
stances of our society. From this will come the best opportunity for us to see how 
what we want to do differs from what we will have to do. Then, appropriate 
choices can be made with a more accurate understanding of their consequences. 

Another set of numbers illustrates the differences in petroleum-resource con- 
centration and their international implications. Norbert Tiemann, administrator 
for the Federal Highway Administration, says in several speeches : "The United 
States oil demand is now about 17 million barrels a day — and growing. However, 
to meet the need, we are producing only some ten million barrels of oil per day 
from half a million wells, an average of about 20 barrels per day per well. In 
contrast, Saudi Arabia could easily produce an equal amount of oil, if it chose, 
from about 700 wells — an average of more than 15,000 barrels a day per well." 
Put another way, the Saudi Arabians can produce with 700 wells what we need 
500,000 wells to produce. 

ENERGY AND COMPETITION 

The conditions of changing and diminishing net energy or energy-investment 
ratio is a fundamental element in all competition, particularly among the nations 
in this turbulent world. Odum and his colleagues are using energy systems 
analysis to ask questions about the competitive position — the leverage — among 
nations. He is concerned that all of the available primary energies available to the 
United States will yield less net energy than Arab oil. This means, he says, that 
we cannot compete using our own resources. 

Odum's analysis, from energy systems studies of net energy done by himself 
and colleagues at the University of Florida, looks at several of our "promising" 
energy technologies. 

For strip mining western coal, we get approximately 3 units of energy back 
from each unit of energy invested. 

For nuclear energy, if the plants last 40 years, we get back between 2 and 3 
units of energy for each unit invested. 

For Arab oil at $10.00 per barrel, we get back about 6.5 units of energy for each 
unit we invest. 

In other words, all the energies readily available to the United States will 
yield less net energy than Arab oil. Again the numbers are approximate. They 
should be seen not as precise statements, but as accurate indications of ques- 
tions we should be asking if, as the numbers portend, we are headed for some 
unpleasant surprises in the near future. 

Will West Germany and Japan be in a better position to outcompete us by not 
having the promised potential of domestic energies to confuse them? They are 
forced to use Arab oil and deal with the balance-of -payment problems. We should 
not forget — oil is of little value if there is no viable industrial society to burn it. 
And what about Russia? What is the net-energy profile of that nation, given 
that its vast resources are spread thousands of miles from industrial centers? 
What questions about competition are exposed when this nation's administration 
makes antique gunboat diplomacy noises in a nuclear missile age? What about 
changes in our global military posture since the 1940's and 1950's, when the 
United States controlled over half of the world's energy? How well can we 
power our threats when we now use 30 percent of the world's energy, but are 
largely dependent on the import of energy and other resources critical to the 
health of our industrial nation? These are all critical questions: who is asking 
them in context with the larger systems in which we live? 

NET ENERGY POLICY 

The concept of net energy cannot be found in the final report of the Ford Foun- 
dation Energy Policy Project. Furthermore, its implication that our economy 
may be uncoupled from energy as reported in the Congressional Quarterly re- 
cently is simply not true. To his credit. David Freeman, director of the Energy 
Policy Project, did invent the phrase "Burn America First" to fault administra- 
tions for their tendency to strip the nation of its remaining diminishing energv 
reserves, thereby potentially threatening even more our future competitiveness. 
The point made by Freeman is not addressed in the administration's Project 



45 

Independence report. Project Independence also failed to recognize the concept 
of net energy, even though Federal Energy Administration officials were quoted 
last summer in Business Week as saying it was a viable concept. 

When Senator Hatfield and Representative McCormick opened the debate on 
net energy, it was no longer an academic or philosophical question. "Burn 
America First" is the announced policy of the Ford Administration and many 
other national leaders. 

ENERGY AND THE INDIVIDUAL 

We tend to look at articles like this one and view them as though we were 
some third party, remote from the reality they attempt to describe. This objec- 
tivity is necessary and wise. What is different here and today, more than ever 
before, is the topic under discussion. It affects you directly — now and into the 
foreseeable future. For example, the question of whether we do or do not get 
pay raises this year or next are very real at this time. This paper ties basic 
influences on inflation directly to changes in energy availability. With a 15-per- 
cent inflation rate last year, you lost that much purchasing power. Did your 
raise cover this loss? Remember, we are talking about energy as it relates to all 
prices. The increased electric and gasoline and fuel prices you are paying are 
only the beginning of a long chain of events that ultimately affects everything 
in our society. 

The possibility of a 20-percent inflation rate is too real to ignore. With it, we 
lose one half of our purchasing power over the next five years. Even at a rate of 
inflation of 15 percent the "half life" of the dollar does not give us too much 
longer. Finally, if you did get a pay raise and produced no more than you did 
last year, your increase is a direct contribution to inflation. 

Individuals must no longer see these questions as separate from themselves. 
We are caught right dead in the center. This begs the question what can we 
do. To ask questions is only a part of the answer. We must ask better questions, 
using processes that allow us to see ourselves in the extended and integrated 
systems of man in nature. As Walter Lippman said, "Our conventional wisdom 
is no longer working ... it was made for a simpler time." 

H. T. Odum would say what wisdom we have made ~by a simpler time. New 
times call for a new wisdom, one he and others feel is available through energy 
systems analysis and a better understanding of man in nature. New perceptual 
processes are at hand, not only with Odum at the University of Florida. We also 
find Jay Forrester at MIT, Dennis Meadows at Dartmouth, K. E. F. Watt at the 
University of California at Davis, and others using new processes to see our 
world. These new processes of perception allowed us to raise the questions found 
on these pages. The universities are exploring new ways to present these proc- 
esses and the questions they ask. You can do your part by seeing that these 
processes and questions, or better ones if needed, are applied in ways that you 
can understand by leaders that we can all trust. 

EPILOGUE 

Homer Lea. a political military strategist, said the following at the turn of 
the century : "To free a nation from error is to enlighten the individual, and 
only to the degree that the individual will be receptive of truth can a nation be 
free from that vanity which ends with national ruin. . . . Xo state is ever 
destroyed except through those avertable conditions that mankind dreads to 
contemplate. Yet nations prefer to perish rather than to master the single lesson 
taught by the washing away of those that have gone before them. In their 
indifference and in the valor of ignorance, they depart together with their monu- 
ments and their constitutions. . . ." 

Past words and actions reflecting "the valor of ignorance" haunt industrial 
society at each step along our journey into the future. In spite of this, there is 
promise in our society's heritage : The "age of power," which began with the 
Industrial Revolution and has culminated in the "communications revolution," 
has, in its passing, provided us with the tools for transcending our ignorance. 
The wisdom of our "simpler times" can be the foundation for a new wisdom. 
The quality of our new existence will be the result of bow well we apply our 
potential to be truly wise to the problems of the great transformation which is 
already upon us. 



46 

Tom Robertson is coordinator of the Energy Center at the University of 
Florida. The center's activities are directed toward better understanding the 
complex interactions between energy and society. An important aspect of this 
goal is the work of Howard Odum and his colleagues, whose energy systems 
analysis, or "energetics," led to the concept of net energy. The purpose of this 
work is to facilitate our transition to an age of increasing shortages by showing 
how we can best utilize the resources that remain. Tom McCall, former governor 
of Oregon, has recently announced that he will join the center staff as head of 
its Institute of Applied Energetics. The institute will encourage the increased 
use of energetics as a tool for understanding the problems of our age, and will 
promote the implementation of policies based on that understanding. 



The Old Economics Has Failed : a New System is Needed to Find the 
True Cost of Energy 

(By Wade Rowland) 

A photographer friend of mine who works for the Toronto Sun once had a 
batch of calling cards printed with the words : 'Colour is a crutch.' I suppose that, 
as an expert in the techniques of capturing volumes of meaning in a single 
black and white news photo, he meant that, while colour can be effective in am- 
plifying the impact of an image, it is an unnecessary frill that at times can get 
in the way of the meaning of thp photograph. 

The card reminded me of a lapel button I used to have (actually, it was kind 
of an anti-button button) that went a little further into metaphysics, asserting. 
'Reality is a crutch.' I have no id°a what the author of that slogan meant by it, 
if anything, but I have lately been thinking that it is a rather neat way of pointing 
out that 'reality' as we have agreed to define it is really just a series of metaphors 
for what goes on in the world around us. We employ the metaphor because only 
a very few of us are able to understand the workings of the universe in the 
more profoundly revealing and truthful way of, say, the mystic. But in the 
same way as colour can get in the way of the meaning of a great photograph — 
particularly of one that records human activity— 'reality' can and does get in 
the way of our understanding how the world works. 

When my photographer friend forgets that colour is only a crutch and gets 
carried away with its inherent power and beauty, nothing much is at risk ; the 
worst that can happen is that he winds up with a bad photograph. But when an 
entire civilization gets carried away by the intrinsic logic and elegance of the 
metaphors it has built up to explain a mechanism much more profound, when 
it becomes so immersed in them that it forgets that they are only metaphors, 
then the survival of that civilization can be at risk. 

That, it seems to me, is what has happened to the industrial civilization of 
the twentieth century, particularlv with regard to the metaphor of modern 
economic theory. 

The history of economics and the elevation of the discipline to the level of a 
pseudo-science is revealing. Orieinallv conceived as a way of helping shopkeepers, 
traders, and factory owners keep track of their money, this rather simple alle- 
gory in dollars and cents has over the past sixty or seventy years groped out to 
subsume all of society's accounting, both social and economic. How we live, 
where we live, what we eat. how we take our pleasures, how we relate to our 
fellow man. are all to a greater rather than lesser extent determined by 'economic' 
considerations, and the sanctions against acting in anti-economic wavs are 
powerful. The sacrifices one must make to live outside the system, to be a free- 
lancer, are not inconsiderable. 

At the same time as this invasion of the remotest crannies of the social 
system was going on. economists were purposefully expunging from their dis- 
cipline every trace of such 'unscientific' concerns as morals and ethics. The 
products of one of the more restrictive areas of an education system so involved 
with the virtues of specialization that it turned out class after class of superbly 
programmed Philistines, few modern economist^ realized that, in chucking out 
moral and ethical knowledge, they were denying their discipline the insights of 
history's greatest and most incisive minds — wisdom that resulted not from the 
structured thought processes of the scientific method, but from long experience in 



47 

the world punctuated by flashes of brilliance and insights sometimes staggering 
in their profundity. 

The results of this systematic avoidance of the richest trove of human under- 
standing is evident to anyone who reads a newspaper. Famines, endemic mental 
disorders, the squalor of our daily lives, pollution disasters, crumbling cities, 
the energy and resource crises, are all symptomatic of a civilization in turmoil 
bordering on chaos. If it was not understood before, the events of the past few 
years should by now have made it abundantly clear that the moral codes held 
in common by the world's great religious philosophies — the insistence that we 
live without greed or envy and that we act without rapaciousness or overween- 
ing pride — are an accurate scientific description of the means of survival on 
this planet. 

That modern economic systems are incompatible with these codes, and delib- 
erately so, we have as evidence the words of Lord Keynes himself, the father 
of modern economic thought. In 1930 he speculated on 'the economic possibilities 
for our grandchildren,' concluding that one day everybody would be wealthy. 
Perhaps then, he said, people will 'once more value ends above means and prefer 
the good to the useful.' 

'But beware,' he continued, "the time for all this is not yet. For at least 
another hundred years we must pretend to ourselves and to every one that fair is 
foul and foul is fair ; for foul is useful and fair is not. Avarice and usury and 
precaution must be our gods for a little longer still. For only they can lead us 
out of the tunnel of economic necessity into daylight.' 

That Keynes was a sensitive and educated man is evident from his biographies, 
so how he came to make such a stupendous error in judgment is a matter for 
conjecture. The fact is, however, that by following his admonition that 'foul 
is useful and fair is not' we have reached the point where there is ground for 
serious doubt as to whether our civilization will still be intact by the time we 
reach his 2030 millenium. 

What we are in need of, then, is a replacement for our present system of eco- 
nomic thought ; a replacement that will not be at odds with our richest fund of 
wisdom and so will not lead us into renewed conflict with our natural environ- 
ment. In other words, what is needed is a new economic system that will op- 
erate within the house rules of the planet as we understand them ; in accord- 
ance with rather than in opposition to the laws of ecology. We need an 'eco- 
e^onomics.' And if it is to be a practical and useful system, it must have a cur- 
rency — a 'dollars and cents' analogue — in which all transactions can be computed 
and accounted for. 

Once it is understood that current economic doctrine was not handed down to 
us on stone tablets, but is simply a metaphorical explanation of our transac- 
tions with nature and with each other, and a faulty one at that, the task of 
deliberately developing a more honest or appropriate system of accounting — a 
new metaphor — seems a less formidable prospect. 

It would appear, in fact, that we are already well on the way to having such a 
svstem, thanks mainly to the work in recent years of Howard Odum, an ecolo- 
gist and professor at the University of Florida, and Joel Schatz, a systems an- 
alyst who was until recently head of the state of Oregon's Office of Energy 
Research and Planning and is now a consultant to several governments. They 
call the new para-economic discipline 'energetics,' and its currency is energy, the 
universal lowest common denominator. Accounting is done in joules or bttjs, 
which can later be translated into francs or dollars or pounds if necessary. 

At the very core of energetics are the laws of nature that govern the function- 
ing of the global ecosystem, the most fundamental of which states that, in terms 
of energy, the cost of any biological or economic enterprise is always greater 
than the benefits. (In physics, this is known as the second law of thermodynamics 
and the 'costs' are measured in terms of increasing entropy.) These processes, 
in other words, only transfrom valuable natural resources (low entropy) into 
waste (high entropy). A second law is derived from this, and it states that long- 
term survival in an ecosystem is possible only for those organisms (man in- 
cluded) that stabilize their growth at a level at which energy consumed is no 
greater than energy available from incoming, renewable sources. That is, you 
won't stay in business long if you live off your capital and throw away your 
income. 

That there are impliactions for energy policy in an accounting system that uses 
energy as its currency is obvious. But before we get into those implications, here 
is what Dr. Schatz has to say about energetics as a policy-formulating tool 
throughout the whole range of economic and social concerns : 



48 

ENERGY INPUT REQUIREMENTS FOR MAJOR CURRENT AND PROPOSED U.S. ENERGY DELIVERY SYSTEMS, NOR- 
MALIZED TO 1,000 BTU DELIVERED ENERGY AND RANKED, WITHIN CLASS, IN ORDER OF EXTERNAL RESOURCE 
REQUIRED 



Energy delivery system 



Primary External External 

Delivered resource subsidy resource 

energy » required » required * required * 

(Btu) (Btu) (Btu) (Btu) 



NONELECTRIC 

Domestic natural gas... 

Liquified natural gas from North Slope. 

Coal gasification (strip mined) 

Solar-space heat... 

Domestic on-shore petroleum 

Geothermal hot water heating 

Alaska North Slope petroleum 

High grade oil shale 

ELECTRIC 
Hydro-electric 

Geothermal steam-electric 

Nuclear fission-electric (LWR). 

Coal gasification-electric (strip mined). 

Coal fired-electric (strip mined) 

Shale oil-electric 



1,000 


1,161 


17 


24 


1,000 


1,287 


33 


48 


1,000 


2,052 


103 


138 


1,000 


3,333 


109 


191 


1,000 


1,093 


186 


250 


1,000 


2,065 


155 


255 


1,000 


1,104 


258 


345 


1,000 


1,238 


359 


483 


1,000 


1,399 


23 


32 


1,000 


7,050 


42 


57 


1,000 


7,425 


213 


451 


1,000 


6,116 


384 


519 


1,000 


3,498 


426 


566 


1,000 


3,692 


1,172 


1,576 



1 Delivered energy: The output energy of a given energy system, e.g., electrical energy at the point and in the form sold 
and transmitted to a commercial, industrial, or residential customer. 

» Primary resource required: The energy potential of the in situ (pre-extraction) primary resource necessary to yield 
1,000 Btu delivered. 

* External subsidy required: The total direct and embodied energy required from other energy systems to build, operate, 
and maintain a given energy system. This external subsidy does not merge with the primary resource flow; it impels it. 
This subsidy is tapped off the output of other energy delivery systems, mostly coal, gas, and oil. 

« External resource required: The energy potential of the in situ (pre-extraction) resource necessary to supply the 
external subsidy which impels the primary resource flow at the 1,000 Btu level of delivery. 



Source: State of Oregon Office of Energy Research and Planning (1974). 



Briefly, energetics is an analytic framework for exploring the role energy plays 
in any natural or social system. It is designed to show how information and 
energy are related in the real world. (What we call 'information' is simply the 
labels and logic that we humans attach to particular patterns of energy in order 
to provide ourselves with an organizational scheme and meaning-structure for 
our sensory-cognitive impressions.) It is rooted in the assumption that all natu- 
ral and social phenomena are fundamental manifestations of energy states and 
transformations. Energy, thus, can be viewed as a common denominator of mean- 
ing which, like an eternal thread, weaves its way through all levels of reality. 

For any specific time period, physical space, and system, there is a unique con- 
figuration, or flow, of energy. This flow of energy can be mapped and measured 
quantitatively in the special linguistic notation of energetics. Such an energy may 
constitute, in effect, a single conceptual lens, or focus, through which substan- 
tially different types of information (apples and oranges) can be organized and 
interrelated. 

A basic computer program or 'energy map' for the state of Oregon has been de- 
veloped at the University of Florida and should be functioning by mid-summer 
this year. It is so comprehensive that it will account for such energy costs as 
food advertising and goods stolen by shoplifters. It should be of enormous value 
to policymakers faced with decisions such as whether and to what extent 
tourism should be promoted, whether a proposed highway or expressway should 
be built, what kind of industry should be encouraged to locate in the state, and so 
on. 

For if the 'public interest' can be said to lie in living within the house rules of 
the planet (or a particular region of it), the value of a system of accounting that 
allows us to see clearly and precisely how our social and economic transactions 
relate to one another and to those limits is to the policy-maker as the telescope is 
to the astronomer and the electron microscope to the biologist. 

A brief, freehand energetic analysis of the current fossil fuel situation provides 
a good case in point. Conventional economic wisdom has it that the current en- 
ergy crisis has resulted not from any ultimate shortage of reserves, but from a 
breakdown in the functioning of the market ; that if more money is channelled 
into exploration and extraction the problem will correct itself. As the price of 
fossil fuels rises, economists tell us, it will become feasible to explore for and 
open up new sources. 



49 

The flaw in this logic lies in the fact that energy in the ground isn't of much 
use until it has been discovered, extracted, processed, and transported to the 
market. It takes energy to do all of these things — energy either from the source 
being tapped or, more often, from some other more easily accessible reserve. 

As we all know by now, energy reserves are becoming more and more remote 
and diffuse, and that means that we are having to funnel more and more energy 
into the energy-getting process. For each successive thousand btus of usable or 
net energy we acquire, we must supply a bigger and bigger energy-getting sub- 
sidy, a subsidy that must be derived from increasingly scarce reserves and one 
that carries with it its own steadily increasing energy-getting costs. 

Figures compiled by the Oregon energy research office show that to deliver 
1,000 btus of electricity supplied by a thermal generator fuelled by strip-mined 
coal, it is necessary to have a reserve of 6,116 btus of coal in the ground (much 
is lost in the extraction process), and the external subsidy required amounts to 
384 btus, for which one must have a reserve in the ground of 519 btus. If the 
generator is to be supplied by oil extracted from high-grade oil shale, the external 
subsidy alone is a staggering 1,172 btus on each 1,000 btus of electricity de- 
livered (see chart). 

In the light of energetics, then, the energy crisis takes on a different appear- 
ance: for one thing, hard-to-get-at reserves like oil shale can no longer be 
thought of as an ace in the hole for providing electricity. (It seems probable that 
the energy subsidy required of tar sand oil would be only marginally lower.) And 
for another, our net energy reserves — the only ones that count — are much smal- 
ler than the gross energy reserve figures quoted daily in the news media. And they 
are growing smaller each day at a rate even greater than simple increases in 
consumption would indicate, since energy-getting costs are steadily increasing. 
Exactly how much smaller net reserves really are can and should be determined 
immediately. 

It also becomes clear from this that so long as we remain in a declining net 
energy situation where it is necessary to spend more and more for each thous- 
and btus delivered, inflation will continue to affect the prices of products in 
which energy is a major cost of production and/or distribution. In a highly in- 
dustrialized society such as ours, that includes just about everything. (The same 
logic, incidentally, applies to other scarce resources such as copper or tin, in 
cases where more plentiful materials cannot be used as substitutes and where 
recycling is not a significant factor.) The only way to end inflation is to adjust 
our continuing energy demands to the point where they can be served mainly 
by safe, continuously renewable energy sources — solar, geothermal, wind, or 
tidal. That is the only way out of the declining net energy trap. It follows that 
the only sane way to be using our conventional fossil fuel reserves at present is 
to be investing them in developing technology and infrastructure that will be 
necessary for large-scale exploitation of the constantly renewable sources. The 
conversion is going to take a lot of energy and it simply won't be available if we 
continue to squander what we have. Remember, our reserves of usable fossil 
fuel are nowhere near as big as we thought they were. 

What is the role of nuclear power in this situation? We have had it drummed 
into us that atomic fission will provide us with a virtually inexhausible source 
of energy, and that would seem to indicate that it provides an escape from the 
declining net energy quandry. 

Unfortunately, this is not the case. Leaving aside the very serious questions of 
safety and waste disposal, questions that may very well curb the expansion of 
nuclear power unless answers are found quickly, energetics makes it clear that 
there is virtually no hope of nuclear energy filling the rising gap between energy 
demand and fossil fuel availability. The problem is the by-now-familiar one of 
increasing energy-getting costs ; the increasing energy costs of building and 
fuelling the reactors. While any single reactor produces about ten times as much 
energy over its lifetime as is consumed during its construction and initial fuelling 
(provided plenty of high-grade uranium ore is available for the fuel), in a rap- 
idly expanding reactor system net energy production may well be zero or worse. 
The reason for this is that for every newly completed reactor that goes on 
stream, several others are still under construction. 

A net energy analysis of an expanding reactor system conducted at Open Uni- 
versity in Britain shows that net energy Incomes available only if the growth rate 
of the system is kept below about 4 per cent a year: i.e. only if the number of 
reactors in the system doubles no more frequently than once every fifteen years. 



50 

As the quality of uranium ore used in fuelling declines, this growth rate must 
also decline if there is to be any net energy production. 

The expansion rate for nuclear energy in Canada projected in official govern- 
ment publications is about 14 per cent a year, for a doubling time of about five 
years. This is clearly a wildly unrealistic expectation, and if it is pursued it will 
worsen, not improve, the energy situation in this country. 

One could go on almost indefinitely examining various sectors of the economy 
from the perspective of energetics (food production and distribution beg for 
attention), but the overriding conclusion to be derived is already clear: stability 
and not growth must become the primary objective of our economic policy-makers. 
This is not to say that all economic growth must come to an end, but rather 
that what growth does take place should occur primarily in sectors whose func- 
tion it is to ease the transition from an economy based on 'foul' waste and expan- 
sion to one based on 'fair' quality and stability. 



It Takes Energy To Get Energy ; the Law of Diminishing Returns 

Is IN Efeeot 

(By Wilson Clark) 

In the mid-19th century, a British company launched the Great Eastern, a coal- 
fired steamship designed to show the prowess of Britain's industrial might. The 
ship, weighing 19,000 tons and equipped with bunkers capable of holding 12,000 
tons of coal, was to voyage to Australia and back without refueling. But it was 
soon discovered that to make the trip the ship would require 75 percent more 
coal than her coal-storage capacity — more coal, in fact, than the weight of the ship 
herself. 

Today the United States is embarking on an effort to become independent 
in energy production, and such a program deserves the kind of analysis that the 
British shipbuilders overlooked. Indeed, our civilization appears to have reached 
a limit similar to that of the Great Eastern : The energy which for so long has 
*T-iven our economy and altered our way of life is becoming scarce, and a number 
of respected experts are suggesting that, without significant changes, our society 
will go the way of the ship that needed more fuel than it could carry. 

In recent years, energy growth in the United States has expanded at a rate of 
nearly four percent per year, resulting in a per capita consumption of all forms 
of energy higher than that of any other nation. U.S. energy consumption in 1970 
was half again as much as all of Western Europe's, even though Europe's popula- 
tion is one-and-a-half times ours. 

As energy consumption has increased in this nation, our energy resources have 
drastically declined. According to M. King Hubbert, a highly respected energy 
and resource expert, the peak for production of all kinds of liquid fossil fuel 
resources (oil and natural gas) was reached in this country in 1970 when 
almost four billion barrels were produced. "The estimated time required to pro- 
duce the middle 80 percent [of the known reserves of this resource]," Hubbert 
says, "is the 61-year period from 1939 to the year 2000, well under a human 
lifespan." 

As available domestic oil and gas resources have declined, we have turned 
more and more to foreign imports — but, since 1973, the price of this essential im- 
ported oil has quadrupled. Recoiling from the specter of another embargo, federal 
officials and industrialists have suggested that the nation develop alternative 
energy sources such as nuclear power, and fossil fuels such as coal and oil shale, 
to bridge the energy gap and enable the nation to become self-sufficient. 

According to John Sawhill, former chief of the Federal Energy Administration, 
"the repercussions of Project Independence will be felt throughout our economy. 
It will have a dramatic impact on the way 211 million Americans work and live." 
The price tag placed on pursuing the energy goals of Project Independence has 
been estimated to fall somewhere between $500 billion and $1 trillion. Raisins: such 
capital for energy development may prove to be the greatest financial undertak- 
ing in the historv of the United States. A srowincr number of experts, however, say 
the eroal of Project Independence may be unreachable. 

The central problem is simply that it takes cnerav to produce new enerpy. In 
other words, in everv process of energv conversion on Earth, some en°rsv i* in OTT i- 
tablv wasted. The laws of thermodynamics, formulated in the last century. 



51 

might be viewed as describing of sort of "energy gravity" in the universe: 
energy constantly moves from hot to cold, from a higher to a lower level. Some 
energy is free for Man's use — but it must be of high quality. Once used, it cannot 
be recycled to produce more power. 

Coal, for example, can be burned in a power plant to produce steam for con- 
version into electric power. But the resulting ashes and waste heat cannot be col- 
lected and burned to produce yet more electricity. The quality of the energy 
in the ashes and heat is not high enough for further such use. 

Numerous studies have indicated that the United States has enormous reserves 
of fossil fuels which can provide centuries of energy for an expanding economy, 
yet few take into account the thermodynamic limitations on mining the fuels left. 
Most cheap and accessible fossil fuel deposits have already been exploited, and the 
energy required to fully exploit the rest may be equal to the energy contained 
in them. What is significant and vital to our future, is the net energy of our fuel 
resources, not the gross energy. Net energy is what is left after the processing, 
concentrating and transporting of energy to consumers is subtracted from the 
gross energy of the resources in the ground. 

Consider the drilling of oil wells. America's first oil well was drilled in Penn- 
sylvania in 1859. From 1860 to 1870, the average depth at which oil was found 
was 300 feet. By 1900, the average find was at 1,000 feet. By 1927, it was 3,000 
feet ; today, it is 6,000 feet. Drilling deeper and deeper into the earth to find scat- 
tered oil deposits requires more and more energy. Think of the energy costs in- 
volved in building the trans-Alaska pipeline (see Smithsonian, October 1974). 
For natural gas, the story is similar. 

Dr. Earl Cook, dean of the College of Geosciences at Texas A. & M. Univer- 
sity, points out that drilling a natural gas well doubles in cost each 3,600 feet. 
Until 1970, he says, all the natural gas found in Texas was no more than 10,000 
feet underground, yet today the gas reserves are found at depths averaging 
20,000 feet and deeper. Drilling a typical well less than a decade ago cost 
$100,000 but now the deeper wells each cost more than $1,000,000 to drill. As oil- 
men move offshore and across the globe in their search for dwindling deposits of 
fossil fuels, financial costs increase, as do the basic energy costs of seeking the 
less concentrated fuel sources. 

Although there is a good deal of oil and natural gas in the ground, the net 
energy — our share — is decreasing constantly. 

The United States has deposits of coal estimated at 3.2 trillion tons, of which 
up to 400 billion tons may be recoverable — enough, some say, to supply this na- 
tion with coal for more than 1,000 years at present rates of energy consumption. 
And since we are dependent on energy in liquid and gaseous form (for such 
w T ork as transportation, home and industrial heating), the energy industries 
and the Federal Energy Administration have proposed that our vast coal de- 
posits be mixed and then converted into gas and liquid fuels. 

Yet the conversion of coal into other forms of energy, such as synthetic natural 
gas, requires not only energy but large quantities of water. In fact, a panel of 
the National Academy of Sciences recently reported that a critical water short- 
age exists in the Western states, where extensive coal deposits are located. 
"Although we conclude that enough water is available for mining and rehabilita- 
tion at most sites," said the scientists, "not enough water exists for large-scale 
conversion of coal to other energy forms {e.g., gasification or steam electric 
power). The potential environmental and social impacts of the use of this water 
for large-scale energy conversion project would exceed by far the anticipated 
impact of mining alone." In fact, the energy and water limitations in the West- 
ern states preclude more than a fraction of the seemingly great U.S. coal deposit? 
from ever being put to use for gasification or liquefaction. 

The prospects for oil shale development are not as optimistic as some official 
predictions portend. Unlike oil, which can be pumped from the ground relatively 
easily and refined into useful products, oil shale is a sedimentary rock which 
contains kerogen, a solid, tarlike organic material. Shale rock must be mined 
and heated in order to release oil from kerogen. The process of mining, heating, 
and processing the oil shale requires so much energy that many experts believe 
that the net energy yield from shale will be negligible. According to Business 
Week, at least one ma.ior oil company has decided that the net energy yield from 
oil shale is so small that they will refuse to bid on federal lands containing 
deposits. And even if a major oil-shale industry were to develop, water supplies 
would be as great a problem as for coal conversion, since the deposits are in 



52 

water-starved Western regions. The twin limiting factors of water and energy 
will preclude the substantial development of these industries. 

Nuclear power is seen as the key to the future, yet an energy assessment of 
the nuclear fuel cycle indicates that the net energy from nuclear power may 
be more limited than the theoretically prodigious energy of the atom has promised. 

Conventional nuclear fission powered plants, which are fueled by uranium, con- 
tribute little more than four percent of the U.S. electricity requirements at pres- 
ent, but according to the Atomic Energy Commission, fission will provide more 
than half of the nation's electricity by the end of the century. Several limitations 
may prevent this from occurring. One is the availability or uranium ore in this 
country for conversion to nuclear fuel. According to the U.S. Geological Survey, 
recoverable uranium resources amount to about 273,000 tons, which will supply 
the nuclear industry only up to the early 1980s. After that, we may well find 
ourselves bargaining for foreign uranium, much as we bargain for foreign 
oil today. 

According to energy consultant E. J. Hoffman, however, an even greater problem 
with nuclear power is that the fuel production process is highly energy-inten- 
sive. "When all energy inputs are considered," he says, such as mining uranium 
ore, enriching nuclear fuel, and fabricating and operating power plants and re- 
processing facilities, "the net electrical yield from fission is very low." Optimistic 
estimates from such sources as the President's Council on Environmental 
Quality say that nuclear fission yields about 12 percent of the energy value of 
the fuel as electricity : Hoffman's estimate is that it yields only 3 percent, That 
advanced reactors might have a higher net yield is one potential, but largely 
unknown at present, since such reactors have not yet been built and operated 
commercially. Other nuclear power processes, such as nuclear fusion, have 
simply not yet been shown to produce electricity, and so they cannot be counted 
upon. Even the more "natural" alternative energy sources, such as solar power, 
wind power and geothermal power, have not been evaluated from the net energy 
standpoint. They hold out great promise — especially from a localized, small- 
scale standpoint. Solar energy, for example, is enormous on a global scale but 
its effect varies from one place to another. However, the net energy yield from 
solar power overall might be low, requiring much energy to build elaborate 
concentrators and heat storage devices necessary. 

What about hydrogen as a replacement fuel? By itself, hydrogen is not at all 
abundant in nature, and other energy sources must first be developed to power 
electrolyzers in order to break down water into hydrogen and oxygen. The energy 
losses inherent in such processes may result in a negligible overall energy yield 
by the time hydrogen is captured, stored and then burned as fuel. An indication 
of the magnitude of this problem has been given by Dr. Derek Gregory of the 
Institute of Gas Technology in Chicago, who points out that to substitute hydro- 
gen fuel fully for the natural gas currently produced would require the con- 
struction of 1,000 enormous one-million-kilowatt capacity electric power plants 
to power electrolyzers — more than twice the present entire installed leectrical 
plant capacity of the nation. 

While much of this kind of analysis is apparently new to most energy planners, 
it also represents more than an analogy to the cost-accounting that is familiar 
to businessmen investing dollars to achieve a net profit. The net energy approach 
might provide a new way of looking at subjects so seemingly disparate as the 
natural world and the economy. 

DOLLAR VALUES OF NATURAL SYSTEMS 

An outspoken proponent of the net energy approach is Dr. Howard T. Odum, 
a systems ecologist at the University of Florida. In the 1950s, Odum analyzed 
the work of researchers trying to grow algae as a cheap source of fuel, and found 
that the energy required to build elaborate facilities and maintain algae cultures 
was greater than the energy yield of the algae when harvested for dry organic 
material. The laboratory experiment was subsidized, not by algae feeding on free 
solar energy — which might have yielded a net energy return — but by "the fossil 
fuel culture through hundreds of dollars spent annually on laboratory equipment 
and services to keep a small number of algae in net yields." 

With his associates at the University of Florida, Odum began to develop a 
symbolic energy language, using computer-modeling techniques, which relates 
energy flows in the natural environment to the energy flows of human technology. 



53 

Odum points out that natural sources of energy — solar radiation, the winds, 
flowing water and energy stored in plants and trees — have been treated as free 
"gifts'' rather than physical energy resources which we can incorporate into our 
economic and environmental thinking. In his energy language, however, a dollar 
value is placed on all sources of energy — whether from the sun or petroleum. 
To produce each dollar in the economy requires energy — for example, to power 
industries. The buying power of the dollar, therefore, can be given an energy 
value. On the average, Odum calculates, the dollar is worth 25,000 calories (kilo- 
calories, or large calories) of energy — the familiar energy equivalent dieters know 
well as food values. Of this figure, 17,000 calories is high-quality energy from 
fossil fuels and 8,000 calories low-quality energy from "natural" sources. In 
other words, the dollar will buy work equal to some-mechanical labor, represented 
by fossil fuel calories, and work done by natural systems and solar energy. 

Odum's concept of energy as the basis of money is not new ; a number of 
19th-century economists thought of money or wealth as deriving from energy 
in nature. The philosophy was expounded earlier in this century by Sir Frederick 
Soddy. the British scientist and Nobel Laureate, who wrote that energy was the 
basis of wealth. "Men in the economic sense," he said, "exist solely by virtue of 
being able to draw on the energy of nature. . . . Wealth, in the economic sense 
of the physical requisites that enable and empower life, is still quite as much 
as of yore the product of the expenditure of energy or work." 

Odum views natural systems as valuable converters and storage devices for 
the solar energy which triggers the life-creating process of photosynthesis. Even 
trees can be given a monetary value for the work they perform, such as air 
purification, prevention of soil erosion, cooling properties, holding ground water, 
and so on. In certain locations, he says, an acre of trees left in the natural state 
is worth more than $10,000 per year or more than $1 million over a hundred-year 
period, not counting inflation. Last year, he calculated that solar energy, in 
conjunction with winds, tides and natural ecological systems in the state of 
Florida, contributed a value of $3 billion to the state, compared to fossil fuel 
purchases by the state's citizens of $18 billion per year. 

The value of the natural systems to the state had never before been calculated. 
"These parts of the basis of our life," says Odum, "continue year after year, 
diminished however, when ecological lands that receive sun, winds, waves and 
rain are diverted to other use." He is now developing a "carrying capacity" plan 
for the future development of the state which has attracted* the interest of the 
state legislature. 

Odum's work may lead to eminently practical applications, by indicating direc- 
tions in which our society can make the best use of energy sources and environ- 
mental planning. One application is to use natural systems for treating wastes, 
rather than using fossil fuels to run conventional waste-treatment plants. "There 
are." he says, "ecosystems capable of using and recycling wastes as a partner 
of the city without drain on the scarce fossil fuels. Soils take up carbon mon- 
oxide, forests absorb nutrients, swamps accept and regulate floodwaters." He 
is currently involved in a three-year program in southern Florida to test the 
capability of swamps to treat wastes, and demonstrate their value to human 
civilization as a natural "power plant." The work, supported by the Rockefeller 
Foundation and the National Science Foundation, has drawn the attention and 
interest of many community and state governments. 

According to Odum's energy concepts, a primary cause of inflation in this 
country and others is the pursuit of high economic growth with ever-more costly 
fossil fuels and other energy sources. As we dig deeper in our search for less- 
concentrated energy supplies to fuel our economy, the actual value of our cur- 
rency is lessening. "Because so much energy has to go immediately into the 
energy-getting process," he notes, "then the real work to society per unit of 
money is less." 

Economists, who generally resent intruders on their turf, have not embraced 
this equation of energy and money with much enthusiasm, but it is gaining 
adherents in several quarters. According to Joel Schatz of Oregon's energy plan- 
ning office, Odum's work leads the way toward effective government planning in 
this age of economic uncertainty. "The more successful the United States is in 
maintaining or increasing its total energy consumption," he says, "under con- 
ditions of declining net energy, the more rapidly inflation, unemployment and 
general economic instability will increase." Many people currently consider this 
disruption only an economic crisis, says Schatz. rather than what he believes 
it really is: a symptom of a continuing and deepening energy crisis. 



54 

There are signs that the net energy approach is being taken seriously even by 
the architects of Project Independence. Eric Zausner of the Federal Energy 
Administration says that net energy is a "useful concept" which is under investi- 
gation. "Net energy flows," he adds, "have practical implications in the new 
and exotic fuels, such as oil shale. With coal, there is no issue, since there is a 
net output of energy. But some of the new processes, such as shale oil processing 
in situ, net energy flow is a very important consideration in whether we should 
do it or not." 

TO THE CREDIT OF LIVE TREE POSTS 

Present-day ecologists are by no means the first to see the value — the dollar 
value — in employing natural systems to do work for Man. Discussing the intro- 
duction of wire for fencing into the United States, Eric Sloane, in his book 
Reverence for Wood, mentions an address that was read before the Philadelphia 
Agricultural Society on January 2, 1816. 

The 1816 account spoke of 'living trees connected with rails of wire,' and true 
to the early American philosophy of looking far ahead, it compared the cost of 
wire fencing with wood fencing over a period of fifty years. It came to the 
conclusion that there was a cash saving of $1,329 per hundred acres enclosed. 
The plan, however, was indeed unique for it enabled the fence to earn money! 
Why plant dead posts in the ground and wait for them to rot? Why not plant 
live trees instead and let them bear fruit and nuts and firewood which would 
then give profit to the farmer ? Using a hundred acres as an example, the Society 
suggested the following plan of live tree posts and showed what they might earn 
a farmer within fifty years (allowing no harvest for the first ten years of 
growth) : 

244 apple trees producing $1 per year $244 

30 cherry trees producing 504 per year 15 

20 pear trees producing 50<f per year 10 

10 plum trees producing 3 

10 shellbark trees producing 10 

50 chestnut trees producing 12 

5 butternut trees producing 20 

5 English walnuts producing 5 

20 walnut trees producing 5 

250 buttonwood trees (24 cords firewood taken from tops) 72 

Multiplied by 40 years' harvests 15, 840, 396 

Deduct the cost of wire rails 1, 751 

To the credit of live tree posts and wire fence in 50 years 14, 089 

Congressman George Brown, Jr., a physicist from Southern California and 
one of a bare handful of scientifically trained members of Congress, goes much 
further. He believes that the new Office of Technology Assessment in the Congress 
should undertake a broad energy analysis, encompassing the net energy approach, 
of the widespread implications of the administration's plans for Project Inde- 
pendence. "We must start with the assumption that the energy available to do 
work is declining. This one assumption, which is firmly based on the laws of 
physics, will revolutionize economic policy once its truth becomes known. . . . 
The implications of the limits to growth of our economic systems are just begin- 
ning to be understood," says Congressman Brown, pointing out that the net 
energy approach indicates the inevitability of a national shift of emphasis toward 
a steady-state economy. "While this view is not yet widely held in Congress, 
the ranks of advocates are growing." 

Since the Industrial Revolution, the Western world has been engaged in a 
srreat enterprise — the building of a highly complicated technological civilization. 
The Western "growth" economy (which today also characterizes Japan) has 
been made possible by seemingly endless supplies of inexpensive energy. One im- 
plication of the net energy approach is that a vigorous and wide-reaching con- 
servation program may be the only palliative for inflation. 

Another implication is that the days of high growth may be over sooner than 
most observers have previously thought. For it is increasingly apparent that 
today's energy crisis is pushing us toward a "steady-state" economy : No oup 
yet knows what such an economy will look like or what social changes will result. 
But it would seem to be about time to start thinking seriously about it. 



55 

Energy Analysis and Public Policy 

the enebgy unit measures environmental consequences, economic costs, 
material needs, and resource availability 

[By Martha W. Gilliland] 

Responsible development of energy resources and allocation of energy research 
and development monies requires an analysis of many social, economic, and 
environmental options. Technology assessment and the environmental impact 
statement have evolved as mechanisms through which options can be identified, 
analyzed, compared, and subjected to public scrutiny. Both mechanisms require 
the analyst to consider potential impacts ranging from those which can be 
rigorously quantified to those which are inherently nonquantifiable. 

A major difficulty, one that exacerbates the uncertainty with which decision- 
makers are almost always confronted, is that different units are commonly used 
in measuring impacts. One of the most commonly used units is dollars. Economists 
often use sophisticated techniques to convert a broad range of "apples and 
oranges" impacts into dollars. Environmental impacts are typically treated as 
externalities and stated in dollar amounts. But this attempt to evaluate all, or 
even most, impacts in terms of dollars is being challenged. A growing number 
of ecologists and environmental interest groups argue that dollars are an inappro- 
priate measure for some impacts and that economic estimates of impacts repre- 
sent, at best, only a fraction of the true environmental costs or benefits. 

An example of the inadequacy of dollars as an assessment measure is the 
mineral resource classification system, utilized within the Department of the 
Interior by the Bureau of Mines and the U.S. Geological Survey. In an attempt 
to provide realistic energy estimates, Interior's classification system subdivides 
resources according to two criteria : the extent of geological knowledge about 
the resource, and the economic feasibility of its recovery. Reserves generally 
refer to economically recoverable material in identified deposits, whereas 
resources include deposits that cannot be recovered due to economic and legal 
constraints (1). However, definition of reserves using an economic criterion 
carries an implicit bias. At best the criterion provides information on whether 
or not the costs of bringing the resource to the consumer are competitive with 
the costs for resources already in production ; thus, the reserve estimates change 
every year and yield little insight into quantities available for the long term. 

What is needed to improve the analysis of interrelations and trade offs among 
environmental consequences, economic costs, material requirements, and resource 
availability is a comprehensive but simplified set of consistent measures drawn 
from a single external conceptual system. The energy accounting procedures 
or net energy analysis utilized by Odum (2), Berry and Fels (3), Chapman (4), 
and Slesser (5) provide such a mechanism. 

The remainder of this article is divided into three parts, (i) The concept of 
net energy is discussed, including a description of the means by which net 
energy is measured, its relationship to energy demand, material shortages, dollar 
costs, environmental stress, and reserve estimates: (ii) net energy analysis is 
demonstrated through an evaluation of geothermal energy development ; and 
(iii) some observations are made concerning the uses and limitations of the 
technique in the public policy-making process. 

NET ENERGY AND ENERGY SUBSIDY 

Net energy has been defined as the amount of energy that remains for con- 
sumer use after the costs of finding, producing, upgrading, and delivering the 
energy have been paid (2). In Fig. 1, these energy costs are conceptualized and 
illustrated as energy subsidies, or feedbacks of high-quality energy which serve 
to "open the valves" for development of more energy. Indications are that, as 
we extract more dilute, deeper, and dirtier energy sources, the energy subsidy 
required to extract and upgrade the new sources increases. Some portion of each 
year's new energy demand represents additional subsidies to energy extraction. 
Consequently, an increase in energy demand or consumption may not represent 
an increase in the amount of energy available to do work in the consuming 
sectors of society. The entire increase may be required to get the new energy. Note 
that this has not always been true, since technological advances sometimes 
compensate for any decrease in the quality of the resource. The introduction 



56 

of solid state electronics into the electronics industry is a case in point. Electric 
power generation is another example. When efficiencies increased and fuel oil 
costs decreased ( due to advances in drilling technology ) , there was a net energy 
increase. Whenever new technological capabilities increase the efficiency of 
performing the same task, net energy increases. These technological advances 
themselves require energy (for research and development) ; however, this energy 
investment has traditionally made large energy savings possible. 

In Fig. 2, the relation between money and energy is illustrated in more detail 
and the external inputs or subsidies are divided into three types ; direct energy, 



Subsidies (S) 




Energy (N) 



R=D+T 
N=D-S 

Physical and 

Thermodynamic 

Losses (T) 

Fig. 1. Functional relationship among net energy, energy demand, and energy subsidy. 




'Society 



Net Energy=0-(S|»S 2 »S3) 
Net Energy Rotco = ^ ^ t ^ 



Physical and 

Thermodynamic 

Losses 



Figure 2. — Categorization of energy subsidy types and the countercurrent rela : 
tion of dollar flow to energy flow. Solid lines represent energy flow and dashed 
lines represent dollar flow. 



57 

material, and environmental subsidies. The processed energy used for process 
heat, in transportation and in manufacturing materials, is a direct energy 
subsidy. 

Material subsidies are less straightforward. They may include goods, services, 
capital, labor, and information. Material, labor, and capital requirements are 
most often measured in terms of economic costs. However, estimates of the energy 
values of these inputs can be made by evaluating the fuels needed for resource 
extraction, transportation, manufacturing, labor, services, and capital expendi- 
tures. From a/n analysis of the network of processes which contribute materials 
to manufacture a commodity, the inputs of the suppliers and of their suppliers 
can be identified. The energy required to manufacture each input can be obtained 
from a number of sources. Data in raw form are given in the Imput-Output 
Structure of the U.S. Economy, and in the Census of Manufacturers (6). In 
addition, several documents now give energy cost data in a more usable form 
for selected materials (3,7, 8). 

It should be noted, however, that procedures for energy accounting are cur- 
rently not consistent, consequently the actual analysis is not as straightforward 
as my description might suggest. For example, some investigators do not give 
labor an energy value at all, others assign it the energy content of the food the 
worker eats, and still others assign to it the total energy consumed by each worker 
(as measured by the goods, services, and food he consumes). Capital depreciation 
is often not included, but some authors assign to it the energy cost of replacing 
the capital goods. Some of the procedures now in use for energy accounting are 
compared and demonsrated in a series of articles in Energy Policy (4, 8). 

When all input requirements are analyzed, it becomes clear that energy limits 
the ability to obtain any input. This had led to the concept of energy as the 
ultimate limiting factor, which is to say: (i) that energy is the only commodity 
for which a substitute cannot be found, (ii) that potential energy is required to 
run every type of system, and (iii) that energy cannot be recycled without violat- 
ing the second law of thermodyamics. 

Project Independence identified many kinds of constraints or limiting factors 
on development of energy resources — shortages of steel, draglines, drilling rigs, 
certain catalysts, water, and certain types of manpower were discussed. In fact, 
however, all of these have a common denominator, energy. With ample energy, 
all materials can be produced or substitute materials found. For example, sea- 
water can be desalinated and pumped to the arid West for oil shale and coal 
development, synthetic substitutes for catalysts can be made, and ash and 
radioactive waste can be rocketed into the solar system. The sulfur can be taken 
out of the coal either before combustion, during combustion, or with stack gas 
cleaning technologies; we can drill to 30,000 feet (9000 meters) for natural gas, 
extract the oil from oil shale and reclaim the land, and recover additional oil from 
oil reservoirs using advanced recovery techniques. However, all of these material 
needs and advanced processes require energy ; thus energy itself is an important 
limiting factor to increasing energy supply. 

The environment also subsidizes energy development, because it provides direct 
services to man. Woodwell (9) refers to these as the "public service functions 
of nature." For example, terrestrial ecosystems purify the air by absorbing and 
recycling air polutants ; similarly, aquatic ecosystems purify the water. Through 
soil stabilization and evapotranspiration ecosystems maintain the hydrologic 
cycle and the quantity of water suplies. They also control the diversity of plant 
and animal populations, provide recreational opportunity, and produce useful 
products such as food and lumber. Recently, pollution has increased to levels 
beyond the absorptive capability of the ecosystem, thereby causing changes in the 
ecosystem (usually toward less productivity). When changes are significant, 
society pays to mitigate the ecosystem damage through "environmental tech- 
nology," that is, stack gas cleaners and advanced waste water treatment plants. 

Dollar evaluations of impacts may account for the cost of the environmental 
technology or the cost of crop damage, but the energy value of the environmental 
subsidy is much larger since the ecosystems deal with lower levels of pollution 
and provide many other services without cost. Schumacher (JO) argues that 
"production depends heavily on the capital provided by nature in the form of air, 
water, and resources," and that "we treat this capital as income, and value it at 
nothing." A dollar evaluation based on the cost of controling polution, providing 
water, recreation, and other services where no ecosystems exist at all might come 
closer to measuring the total environmental subsidy. 



58 

Thus purely natural ecosystems, as well as agricultural systems such as North 
Dakota's wheatlands and Montana's cattlelaeds, have high energy value for 
man. Their value will be lost for some time while coal is stripped from the sub- 
surface. Western oil shale and coal resources are located in water-scarce regions 
and their exploitation consumes large quantities of water. If a decision is made 
to allocate water to western energy development, many agricultural users may 
not only be denied water but the quality of what is available many be reduced. 
Until vegetation is reestablished, runoff will be much greater than on grass- 
covered soils. These losses in natural value must be included as lost subsidies in 
net energy calculations. They represent losses to society that are partially paid 
for with expensive technology and sometimes compensated for by direct payment 
to those receiving the damages, as is now being considered for the coastal 
states adjacent to outer continental shelf oil and gas development. 

The energy value of environmental subsidies is generally evaluated by cal- 
culating the losses in photosynthetic activity (as reflected in reduced gross pri- 
mary productivity) caused by land disruption or ecosystem change (11). Gross 
primary productivity is a measure of the amount of sunlight captured and 
concentrated by plants and, consequently, is a measure of the work the ecosystem 
does. Additional measures may also be important, such as the work the sun does 
by inducing a heat gradient within the ecosystem (measured by the Carnot 
ratio), and the work done by the kinetic energy of the wind or tides (in the 
case of a coastal system) coming from outside the system. If the heat gradient 
within the ecosystem or wind flow through it were changed by the development, 
these changes would also affect the net energy calculation. 

In summing the various types of energy subsidies, all energy measures must be 
of the same quality. Energy forms are the same quality if they are equivalent in 
their ability to do work. For example, a calorie of electricity can do more work 
than a calorie of coal or oil and both can do more w T ork than a calorie of sunlight. 
Energy quality is calculated by evaluating the energy used in converting from one 
energy form to another, that is, by evaluating the amount of one type of energy re- 
quired to develop another. In the conversion of coal to electricity, physical and 
thermo-dynamic losses occur and auxiliary energy is used within the process and 
in maintaining the industry structure. The ratio of energy delivered to the sum of 
the losses plus the auxiliary energy is the quality conversion factor for coal to elec- 
tricity. As such, 3% units of energy in the form of coal are equivalent to 1 unit of 
electricity in their potential to do work (12). Most people are familiar with 
quality differences between electricity and coal, but there are similar differences 
among other energy forms. For example, petroleum is approximately 2000 
times more concentrated than the sun's energy, 20 times more concentrated than 
photosynthetic energy (sugar), and 40 times more concentrated than wind energy 
(12). Since electricity is 3.5 times more concentrated than petroleum, it requires 
7000 calories of sunlight to produce 1 calorie of electricity (2000X3.5). Higher 
quality energy can do w T ork that was not possible at all with the original energy 
forms ; electronic communication is not possible without electricity ; and, as 
defined here, information is the highest quality energy form, since it requires 
large amounts of time and energy (for research teams, educational institu- 
tions, and libraries) to develop. In order to obtain the total energy subsidy for 
a process, all types of subsidies must first be converted to the same quality. 

MONEY AND ENERGY 

Figure 2 shows the flow of money in the opposite direction to the flow of 
energy, indicating that the mining and processing sectors pay society for material 
and information, and society pays the processing sector for high-quality energy. 
The ratio of the two countercurrent flows (money and energy) is the price of the 
material (dollars per kilocalorie) or energy expended per dollar cost (kilocalories 
per dollar). The average price, or energy expended per dollar, for any given year 
is the ratio of total U.S. energy consumption to gross national product (GNP) for 
that year. In real dollars, this ratio was 21,200 in 1963 ; in 1070 it was 17,300 and 
in 1072 it was 15,800 (13) . With the use of this ratio it is possible to convert dollar 
cost into energy subsidy. However, this represents average dollar to energy con- 
versions for the entire economy, so that only an approximate energy value for 
a wide mixture of goods and services can be obtained. In addition, the dollar 
costs may include hidden institutional subsidies (that is, tax depletion allowance) , 
or represent some regulated price rather than true costs. For specific sectors of the 
economy such as primary metals, mining, and petroleum refining, more accurate 



59 

dollar to energy conversions can be estimated. Kylstra {13) calculated that in 1963 
the primary metal sector used 28,665 kilocalories per dollar while the mining sec- 
tor used 22,050 kilocalories per dollar. Up to date, dollar to energy conversions are 
needed if net energy analyses rely on costs. In principle, however, it is possible to 
account for all the energy subsidies directly without relying on cost and dollar to 
energy conversions. The important point is that a conversion and functional rela- 
tion between the flow of money and the flow of energy exists, with the ratio of 
energy to money decreasing as one progresses within the economy from the fuel 
processing and primary raw materials processing sectors through manufacturing 
and energy conversion processes and Anally to the consumer, who receives the 
smallest amount of energy for his dollar. 

Figure 2 indicates that there is no money flow associated with either environ- 
mental subsidies or raw energy flow. We do not pay nature for each acre of land 
taken out of biological production, nor do we pay nature for the millions of years 
of work it did in making coal or oil. We pay industry to mitigate the environ- 
mental losses through environmental technology and to extract and upgrade the 
coal and oil. As indicated in Fig. 2, money circulates in the economy, but the sun 
and the raw coal, oil, gas, and uramium drive that circulation. 

ECONOMIC FEASIBILITY VERSUS ENERGY FEASIBILITY 

Economic feasibility studies done in the past for extraction of oil from oil shale 
concluded that it was economically unsound, that is, large monetary expenditures 
were required. In terms of Fig. 2, this also means that large energy expenditures 
(labor, materials, water, and capital structure) were required. The amount of 
energy in the feedback loops for oil shale development was larger than for other 
energy sources, and that is what made it uneconomic. Recent economic studies 
concluded that extraction of oil from oil shale may be economically feasible, 
although the amount of energy in the feedback loops has not changed. The change 
is in the fact that other energy sources now require the same amount of subsidy ; 
thus, oil shale now appears to be competitive. The net amount of energy which 
will go to society has not changed either, but where U.S. Geological Survey 
reserve estimates previously indicated zero, they now will show some eco- 
nomically feasible quantity. The true reserve to society is probably neither 
number. Net energy estimates will not change with changing dollar values. They 
will, in fact, remain constant with time unless technological advances in conver- 
sion efficiencies occur. Thus, the economic costs may measure the relative amount 
of energy in the subsidy (assuming hidden dollars in the form of depletion 
allowances are somehow negated), but they do not provide information on when 
the subsidy exceeds the output. 

TABLE 1— ENERGY FLOW FROM THE WELLHEAD TO THE CONSUMER FOR A 100 MW GEOTHERMAL POWERPLANT 
AT 80 PERCENT LOAD FACTOR FOR 30 YR 

Steam-driven turbines 



Resource flow 

At wellhead 

Input to powerplant 

Steam ejector use 

Total generated as electricity 

Auxiliary power use 

Net output of electricity 

Delivered to consumer as electricity! 

•Based on the Geysers, California (20). 

fBased on a 6 percent energy loss from the wellhead to the powerplant and 11 percent efficiency from the wellhead to 
the transmission line (21). 
transmission and distribution loss of 9 percent. 
••Unknown. 

Economically, geothermal energy development now appears to be a viable 
option. Present average investment costs for geothermal power as ,$250 per 
installed kilowatt. However, as high salinity brines, lower temperature fluids, 
and hot dry rock sources are exploited, these investment costs are expected 
to rise to $500 per kilowatt in constant 1973 dollars. The cost rise is a result 



Dry steam 
reservoir* 
(10» kcal) 


Wet steam reser- 
voir—two-stage 
flashedtOOi* kcal) 


116.0 


164.2 


115.0 


154.4 


4.7 


*• 


18.7 


19.0 


.6 


.9 


18.1 


18.1 


16.5 


16.5 



60 

of the low quality (that is, deeper, more dilute, dirtier) nature of these new 
geothermal reservoirs. They will yield no more energy to society in the future 
than they would now. The reason we are not exploiting them now is that they 
require more subsidy (energy feedback) than competing sources do. 

These examples emphasize the importance of answering the question : how 
much of the projected new energy demand for 1985, 1990, and later will be 
expended to increase or maintain net amounts of energy and thus the real 
GNP, and how much is simply the energy subsidy required to obtain and 
upgrade the new dilute energy sources? Estimates of net reserves require 
answers to questions such as at what combination of depth, energy content, and 
sulfur content does coal cost more energy to extract, clean up, and process than it 
yields? Any coal with better characteristics than his "cutoff" combination is 
part of the reserve. At what depth onshore and offshore will oil and natural gas 
be net yielders ? How much heat or chemicals can be pumped into an oil reservoir 
for secondary or tertiary recovery before more energy is being pumped in than is 
in the oil when it gets to the consumer? What is the "cutoff" combination of heat 
content, mineral content, and depth which makes a geothermal reservoir 
a net yielder? The reserve is the amount that exists with better characteristics 
than the cutoff characteristics. It is this net amount which will allow the United 
States to grow economically. Any amount below the net amount will be reqiured 
just to maintain the present state. Within the net reserve category, some energy 
development will require less subsidy than others, thus some will be more 
economic to extract and process than others. 

GEOTHERMAL ENERGY RESERVES 

It has been argued that, since geothermal energy is the natural heat of the 
earth, the geothermal resource is all of the heat in the earth's crust above 
the mean surface temperature or above 15 °C. Since this heat is diffuse, a geo- 
thermal reservoir is said to occur whenever the heat flow from depth is 
one and one-half to five times the world-wide average of 1.5 X 10~ 6 calories per 
square centimeter per second. In addition, it has been postulated that geo- 
thermal energy from dry hot rocks systems is almost limitless, since drilling 
5.5 to 7.5 kilometers under a typical earth temperature gradient of 25 °C per 
kilometer would yield the required 150 °C to 200° C for geothermal power. On 
this basis, Rex and Howell (14) estimate that 40,000,000 megawatt centuries 
of electricity (megawatts of capacity witli a projected life of a century) 
are available by exploiting hot dry rock at less than 10.5 kilometers. 

On the other hand, the volcanic area being investigated for hot dry rock 
in the Jemez Mountains in New Mexico has a temperature gradient of 180° C 
per kilometer, which is 7.2 times the normal temperature gradient of the earth. 
A temperature of 200°C can be reached within 1.2 kilometers. This system could 
be a net energy producer. 

Ideally, the question which should be addressed is: What combination of 
technological efficiencies, heat flow, and depth yields net energy? Unfortunately, 
data are not yet available to do accurate total net reserve calculations. Svstematic 
and consistent compilations of the energy per kilogram required for 'all types 
of goods and services, and the kilograms of raw and manufactured materials 
required for every major piece of equipment are needed. Thus, the analysis 
below is presented both as a methodology for others to use and develop, and 
as a preliminary step in the evaluation of geothermal net energy reserves 
Two power cycles using energy from two types of geothermal reservoirs were 
considered : a dry steam reservoir with steam driving the turbine, and a wet 
steam reservoir with two-stage flashed steam driving the turbine. As more data 
become available, the comparison will be extended to include binary systems 
and total flow impulse turbines using heat from wet steam reservoirs 'and from 
not dry rock reservoirs. 

Table 1 gives the physical and thermodynamic losses of energy as it is trans- 
formed from the enthalpy (heat content) in the steam or hot water at the well- 



61 

head to electricity delivered to the consumer. Output is based on a 100-megawatt 
(net) capacity power plant operating at 80 percent load factor for 30 years. 
Each system delivers 16.5 X 10 12 kilocalories (electric) to the consumer in 30 
years. Wellhead-to-consumer efficiency including electric transmission losses 
for the dry steam system was 14.2 percent, and for the wet steam system was 
10 percent. 

Table 2 lists energy, material, and environmental subsidies for developing and 
operating a 100-megawatt geothermal power system for 30 years. Details of the 
calculations are given in the notes. The exploration value assumes that one 
out of four land areas acquired will be drilled, that one out of four exploratory 
wells drilled will be completed for testing, and that one out of four of these 
completed wells will locate a field of commercial size (15). As geothermal sites 
become more difficult to locate, the exploration subsidy will increase, reducing the 
overall net energy. The extraction subsidy is based on a drilling time of 40 
days per well and 20 days for cementing. It would increase as deeper reservoirs 
are tapped. The subsidy from the environment (measured as a stress on it) 
includes the reduction in gross primary productivity caused by the land require- 
ments of the geothermal field. The geothermal field is assumed to be located in a 
forested area such as northern California where the Geysers field occurs. 

The sum of all subsidies is about 4600 X 10 9 kilocalories for a dry steam 
field and 5400 X 10 9 kilocalories for a blue-type field using a two-stage flashed 
steam-driven turbine. The total delivered energy from the 100-megawatt power 
plant was 16,500 X 10 9 kilocalories (electric) or 57,750 X 10 9 kilocalories 
(equivalent to petroleum in quality) over 30 years at 80 percent load factor. 
Thus, the ratio of energy delivered to energy subsidy was about 12.6 :1 for the 
dry steam field and 10.7 :1 for the brine system. The environmental subsidy 
was low in each case. However, neither the health effects of sulfur emissions 
nor the biogeological effect of subsidence or induced earthquakes on the land- 
scape have been evaluated. In addition, one could argue that indirect environ- 
mental subsidies for extracting the metals used in the materials and for manu- 
facturing those materials should also be included. If we view our economic 
system as one driven by the sun and raw fuels as in Fig. 2, then we should 
include these indirect environmental subsidies just as we have included the 
indirect energy and material subsidies. Calculating from data given by Kylstra 
(13), I estimate that, for every kilocalorie of fossil fuel subsidy, there is an 
additional 0.3 kilocalorie (equivalent to petroleum in quality) of subsidy from 
nature. To my knowledge, no investigators have included this indirect environ- 
mental subsidy in net energy calculations. 

The largest uncertainty in the numercial values given in Table 2 occurs where 
dollar costs were converted to energy units. These were the energy for explora- 
tion, for maintenance materials in the power plant, and for operating the field, 
power plant, and distribution system. These values could varv as much as 25 
percent and that variance would cause the total energv subsidy to varv bv 17 
percent. 

There are a number of configurations for geothermal power systems, each 
of which would result in a different ratio. For example, the electric power 
generation step requires the highest subsidy ; thus it may be more net energy 
efficient to produce and utilize steam directlv for space heating or industrial 
process heat. However, this requires that the users be located in close proximity 
to the geothermal field. Electricity is high-qualitv energv so that the price of 
producing it as measured by the energy subsidies will always be high Some 
combination of depth and enthalpy of the geothermal fluid represents the point 
where as much energy is required to extract it as is produced. This will varv 
Slightly tor each type of proposed electric power generation facility (steam tur- 
hL 11 H S, fi 1, " I i l 1 ! se / urbines - and hea * exchangers). The geothermal reserves should 
eL^gy tfen^y^lsidy thelr ** ^^ ^^ that iS ' the rati <> ° f ™™*« 



62 

TABLE 2.-ENERGY STUDIES REQUIRED FOR THE DEVELOPMENT AND OPERATION OF A 100-MW GEOTHERMAL 
POWER SYSTEM FOR 30 YR, ALL VALUES ARE EQUIVALENT TO PETROLEUM IN ENERGY QUALITY 



Subsidy types 



Dry steam Wet steam 

reservoir* (10« kcal) reservoirf (lO'kcal) 



Exploration (22) 

Extraction and separation (23): 

Fuel 

Construction and maintenance materials 

Transport of materials 

Steam transport (24): 

Construction and maintenance materials 

Transport of materials 

Construction! and operation of the steam field (25) 

Conversion to electricity: 

Construction materials (26) 

Maintenance materials (27) 

Transport of materials (28). 

Construction! and operation of the power plant (29).. 
Transmission and distribution (30): 

Construction and maintenance materials 

Construction! and operation of the transmission lines. 
Environment (31): 

Field site 

Transmission corridor 

Total subsidy 

Total energy delivered to consumer** 

Net energy ratio— delivered energy to subsidy 



50 



50 



135 

135 

5 


150 

150 

6 


25 
3 

140 


35 

4 

185 


570 
25 
70 

160 


1,140 
35 
140 
215 


2,800 
400 


2,800 
400 


35 

35 

4,588 

57, 750 

12.6:1 


50 

35 

5,395 

57, 750 

10.7:1 



* Steam-driven turbine. 

t 2-stage flashed steam-driven turbine. 

i Excluding materials. 

** 16,500 kilo-calories (electric) X 3.5 is 57,750 kilocalories of petroleum equivalents. 

NET ENERGY RATIOS 

The net energy ratio, as defined above and in Fig. 2, does not include physical 
and thermodynamic losses directly, but is the ratio of delivered energy to the 
energy value of material, environmental, and processed energy subsidies. The 
physical and thermodynamic losses are included only in the sense that increased 
efficiences would reduce the losses and increases the delivered energy value. There 
have been several attemps to calculate these net energy ratios for other energy 
systems. Ballantine (16) calculated that the ratio of energy delivered as elec- 
tricity from Northern Great Plains coal (based on 4700 kilocalories per kilo- 
gram) to energy subsidy is 4:1. Based on a lOOO-megawatt light water nuclear 
reactor, Lem (17) calculated the maximum ratio of delivered electricity to energy 
subsidy as 9:1. Oregon's Office of Energy Planning (18) calculated ratios of 60:1 
for domestic natural gas, 7:1 for high-Btu (British thermal unit) gas from coal, 
and 2.8:1 for oil from ol shale (all nonelectrc uses). Although all of these 
ratios represent delivered energy to subsidy and all are expressed in equivalent 
energy qualities, in each case data were incomplete, so that precise comparisons 
are not possible. 

When the price of oil increased, its net energy ratio decreased, resulting in 
inflation. Imported oil at $2 per barrel has a net energy ratio of 30 to 1, while at 
$11 per barrel the ratio is 6 to 1 (16). The real GNP cannot increase unless the 
economy is driven by energy sources that require little energy to extract. The 
purely economic calculations obscure this fact since they include the effects of 
government policy in subsidizing some resources (that is. nuclear) and not 
others. Government energy policy in areas such as outer continental shelf leases 
for oil, onshore leases for coal and geothermal sources, and tax depletion allow- 
ances could be made on the basis of which resources have the highest ratio of 
delivered energy to energy subsidy. And the U.S. Geological Survey could aid 
policymakers by calculating 'reserves on this basis as well as their economical 
recovery. The economics of the reserve estimate will track the net energy ratio. 



NET ENERGY AND PUBLIC POLICY DECISIONS 

Energy analysis has already captured the attention of persons searching for 
better policy analysis tools. Section 5 of the Non-Nuclear Energy Research and 
Development Act of 1974 (PL 93-577) states, as one of the governing principles 
for researching and demonstrating new energy resources, that "the potential for 



63 

production of net energy by the proposed technology at the stage of commercial 
application shall be analyzed and considered in evaluating proposals." In response 
to this legislation, there are several government agencies involved in standardizing 
energy analysis procedures, and in performing some calculations. The Office of 
Energy Policy of the National Science Foundation (NSF) brought together 
energy accounting researchers at a workshop in August 1975. The objective of 
that workshop was to compare and standardize procedures and determine specific 
policy applications for the analyses. The Energy Research and Development 
Administration (ERDA) has stated that it plans to integrate evaluations of the 
net energy contribution of technologies into the national plan for setting energy 
research needs and priorities (19). ERDA's Office of Planning and Analysis is 
expected to have funding responsibility for these studies. In addition, the De- 
partment of the Interior's Office of Research and Development has contracted 
for energy analysis of several technologies. As a result of the legislation and 
agency interest, there is a probability that net energy analysis may come into 
widespread use. It has the potential to improve the input into the decision- 
making process. 

The data and information provided to policy-makers are almost always in- 
complete and conflicting. Energy analysis may not eliminate the incompleteness, 
but it can reduce the conflicting nature of the inputs. I have shown the special role 
that energy plays in driving the flow of money, in allowing for the extraction, 
manufacture, and transportation of materials, and in allowing for substitution of 
different materials for ones in short supply. Since energy is the one commodity 
present in all processes and since there is no substitute for it, using energy as 
the physical measure of environmental and social impacts, of material, capital, 
and manpower requirements, and of reserve quantities reduces the need to com- 
pare or add "apples and oranges." In energy analysis, many environmental and 
social costs and benefits are internalized directly. For example, the energy value of 
the environment is the amount of the sun's energy used by the ecosystem in 
providing services and products, just as the value of a manufactured commodity 
is the amount of fossil fuel used by the machines in making the product. The use 
of the energy unit makes the two comparable. 

Dollar evaluations do not usually internalize environmental costs, such as air 
pollution, or social costs, such as government subsidies in the form of regula- 
tions, taxes, or research. In addition, dollar evaluations often obscure the larger 
scale effects of an action because the dollar costs and benefits accrue to different 
people at different times. Dollar evaluations also change with time due to the 
changing value of money and assumptions concerning, for example, the discount 
rate. For a specific technology, such as the present nuclear fuel cycle and its 
supporting techniques, the energy evaluation will not change with time. 

Energy analysis of alternative energy supply technologies can provide more 
information of a less conflicting nature to policy -makers. Assuming that more and 
better information improves the quality of decisions, then energy analysis can im- 
prove government policies in areas such as managing public energy lands, regu- 
lating gas, oil, and utility rates, providing tax incentives, and establishing re- 
search emphasis. 

REFERENCES AND NOTES 

1. V. E. McKelvey. Am. Sci. 60, 32 (1972). 

2. H. T. Odum, Ambio 2, 220 (1973) . 

3. R. S. Berry and M. F. Fels, Bull. At. Sci. 29, 11 (1973). 

4. P. F. Chapman, Energy Policy 2, 91 (1974) . 

5. M. Slesser, Tech. Assess. 2, 201 (1974). 

6. Department of Commerce, Input-Output Structure of the T.S. Economy: 1963 
(Government Printing Office, Washington, DC, 1909) ; 1967 Census of Manufac- 
turers (Government Printing Office, Washington. D.C., 1971). 

7. A. B. Makhijani and A. .7. Lichtenberg. Environment 14, 10 (1972). 

8. P. F. Chapman, G. Leach, M. Slesser, Energy Policy 2, 231 (1974) ; D. J. 
Wright, ibid. 2, 307 (1974) ; P. F. Chapman, ibid. 3. 47 (1975). 

9. G. M. Woodwell, Nat. Hist. 83, 16 (1974). 

10. E. F. Schumacher, Small Is Beautiful (Harper, New York, 1973), from 
N. Wade, in Science 1S9, 199 (1975). 

11. H. T. Odum, in Thermal Ecology, J. W. Gibbons and R. U. Sharits, Eds. 
(Technical Information Center, Springfield, Virginia. 1974), p. 628. 



64 

12. H. T. Odum, in Simulation of Macroenergctic Models of Environment, Pow- 
er, and Society, H. T. Odum, Ed. (Energy Center, Univ. of Florida, Gainesville, 
1974), p. 5. 

13. C. D. Kylstra, Energy Analysis as a Common Basis for Optimally Combining 
Alan's Activities and Nature (Energy Center, Univ. of Florida, Gainesville, 1974), 
pp. 16, 20. 

14. R. W. Rex and D. J. Howell, in Geothermal Energy: Resources, Production, 
Stimulation, P. Kruger and C. Otte, Eds. (Standard Univ. Press. Stanford, Cali- 
fornia, 1973) , p. 63. 

15. R. Greider, Geothermal Resources Council Bull. 3, 2 (1974). 

16. T. Ballantine, Net Energy Calculations of Northern Great Plains Coal in 
Power Plants, unpublished paper (Environmental Engineering Sciences, Univ. of 
Florida, Gainesville, 1974), pp. 20, 25. 

17. P. N. Lem, thesis, University of Florida (1973), pp. 150-190. 

18. Oregon Office of Energy Research and Planning, Energy Study (Office of 
the Governor, State of Oregon, Salem, 1974), pp. 150-180. 

19. Energy Research and Development Administration. A National Plan for 
Energy Research, Development, and Demonstration: Creating Energy Choices for 
the Future (Government Printing Office, Washington, D.C., 1975), p. X 3. 

20. C. F. Budd, in Geothermal Energy: Resources, Production, Stimulation, P. 
Kruger and C. Otte, Eds. (Stanford Univer. Press, Stanford, California, 1973), 
pp. 129-144 ; J. P. Finney, ibid., pp. 145-162. 

21. A. L. Austin, G. H. Higgens, J. H. Howard, The Total Flow Concept for 
Recovery of Energy from Geothermal Hot Brine Deposits (Lawrence Livermore 
Laboratory, Livermore, California, 1973), pp. 1-37; Federal Administration, 
Project Independence Blueprint-Geothermal Energy (U.S. Government Printing 
Office, Washington, DC., 1974), p. E1-E5. 

22. Cost is estimated at $32 per kilowatt; 15,800 kilocalories per dollars. 
100,000 kilowatts. 

23. Fuel includes that for drill engines, mud pumps, cement pumps, and 
chemical injection ; drilling depth in dry steam field of 1.8 kilometers and in the 
brine field of 1.4 kilometers for production wells and 0.45 kilometers for reinjec- 
tion wells ; 20 wells per 100 megawatt dry steam plant and 26 production wells, 
plus 13 reinjection wells per 100 megawatt brine plant ; each production well 
lasts 10 years ; drill engines operate 530 hours per kilometer drilled, cement pumps 
operate 270 hours per kilometer drilled, mud pumps operate 280 hours per kilo- 
meter drilled ; all engines are 450 kilowatts and require 2030 kilocalories per kilo- 
watt hour. Construction and maintenance materials include well completion 
materials and drilling equipment. For the dry steam field : 60 wells ; 103,500 kilo- 
grams casing per well ; 78,800 kilograms of cement per well ; 2 valves per well ; 
one steam separator. For the brine system ; 78 production wells ; 26 reinjection 
wells ; 85,300 kilograms of casing per production well. 33,300 kilograms of casing 
per reinjection well ; 59,000 kilograms of cement per production well ; 18,200 
kilograms of cement per reinjection well ; three valves and three separators per 
production well ; 16,600 kilocalories per kilogram of steel ; 3400 kilocalories per 
kilogram of cement. Over the 30-year life, 4 derricks, 8 drill engines, 16 mud 
pumps, 8 cement pumps, 2 blowout preventers, 8 chemical injection pumps, and 20 
drill bits per well are consumed. Transport : 3200 kilometers from the Midwest to 
California ; 0.7 kilocalories per kilogram per kilometer. 

24. All wells are located 0.8 kilometer from the power plant ; for the dry steam 
system, 700,000 kilograms of steel for steam lines ; for the brine system, 950,000 
kilograms of steel for production and reinjection well steam lines, all replaced 100 
percent in 30 years ; 16,000 kilocalories per kilogram of steel. Transport : 3200 
kilometers from the Midwest to California ; 0.7 kilocalories per kilogram per 
kilometer. 

25. For labor, taxes, rents, interest, in constructing and operating the field over 
30 years : $22 million at the dry steam field ; $29 mililon at the brine field ; 6400 
kilocalories per dollars. 

26. Construction materials for dry steam turbine-generator, in 10 3 kilograms ; 
aluminum, 45.4; copper, 91; concrete, 22,700; steel, 8720; stainless steel, 118; 
steel forgings, 76.3; other nonferrous metals, 54.5. Energy content: in 10 3 kilo- 
calories per kilogram : Al, 21 ; Cu 31 ; concrete 3.4 ;steel, 16.6 ; stainless steel, 22 ; 
steel forgeings, 76.3 ; other nonferrous metals, 54.5. Energy content in 10 3 kilo- 
calories per kilogram ; Al, 21 ; Cu. 31 ; concrete 3.4 ; steel, 16.6 ; stainless steel, 22 ; 
steel foregoings, 28 ; other nonferrous metals, 20. Fabrication of the turbine- 



65 

generator. 31.8 X 10 6 kilogram materials ; 9500 kilocalories per kilogram. For a 
brine power plant, the turbine is twice as large to accommodate lower pressure 
and temperature steam. 

27. Over 30 years, $2.5 million for materials is needed for repairs at dry steam 
field power plant ; $3.3 million at the brine field ; 10,000 kilocalories per dollar. 

28. Dry steam, 31.8 X 10 a kilograms, brine, 63.6 X 10 8 kilograms ; 3200 kilo- 
meters ; 0.7 kilocalories per kilogram kilometer. 

29. For labor, taxes, rents, interest : $2.5 million for construction. $23 million for 
operation over 30 years at the dry steam power plant ; $3.3 million for construction. 
$30 million for operation over 30 years at the brine power plant ; 6400 kilocalories 
per dollar. 

30. For transmission and distribution, 7.7 mills per kilowatt hour cost; 55 
percent for materials at 10,000 kilocalories per dollar. 5 percent for fuel at 240,000 
kilocalories per dollar, and 40 percent for operating at 6400 kilocalories per 
dollar. 

31. Dry steam field ; 40 hectares for direct use for well site pads, work areas, 
steam lines, and power plant plus 360 hectares indirect use. Brine field : 60 hec- 
tares for direct use plus 540 hectares for indirect use. In 1970. 400.000 megawatts 
of capacity withdrew 1.6 million hectares of land for transmission lines : therefore 
100 megawatts withdraws 400 hectares, of which 20 percent is direct use. Forest 
productivity in northern California is 5000 kilocalories per square meter per year. 
Direct use eliminates productivity ; indirect use reduces it by half for 60 years. 
30 years during field use and 30 years for ecosystem recovery. Energy quality con- 
version factor is 1/20. 

32. I thank I. White, D. Kash, and R. Ryeroft for evaluating the manuscript. 
This article is the result of research carried out within the Science and Public 
Policy Program under NSF grant No. SI A 74-17866. 



Net Energy Analysis Can Be Illuminating 
(By Rice Odell) 

SUMMARY 

Energy analysis is an emerging methodology used to better understand our 
current energy and economic problems and to facilitate the discovery of solu- 
tions. It involves computations of "net energy*' derived from a system — the 
amount of energy remaining when energy inputs have been subtracted from 
energy outputs. It is particularly relevant since we have grown into an indus- 
trial society based on intensive use of energy — and since it is taking more and 
more energy to produce energy. Analytical techniques must be used carefully, 
however. Different forms of energy have different uses and values to society. 
They cannot always be massaged together. It is also difficult to establish the 
boundaries of energy analysis. Energy analysis can be a useful supplement to 
economic analysis. In the long run, however, social constraints may be the most 
obstinate of all. 

For a long time, the chief criteria for decision-making in our society have been 
economic — values defined in dollars, cost-price mechanisms, profits. But now an- 
other measure of value has taken on great importance: energy. 

We are accustomed to asking : How much will it cost, and how great will tbe 
profit be? Now there are other relevant questions : How much energy does it use. 
and how much energy will be saved or returned? 

Spaking of nuclear power projects, W. Kenneth Davis, a vice president of 
Bechtel Power Corp.. has said there is ''an energy investment just as there is a 
financial investment and there is a necessity to show an energy profit." : 

There is now emerging a rather specialized methodology for measuring and 
understanding the energy flows in a given project, process or product. It is 
energy analysis. It is used, for example, to calculate tbe energy inputs and out- 
puts, or the "net energy" derived in a system. It lias been called energy account- 
ing, energy cost evaluation, net energy analysis and other labels. (Last year, at 
a meeting in Sweden, an international group of pioneers in the field agreed to 
stick to the term energy analysis.) 

It is not viewed as a substitute for economic analysis, but as a complement to 
it. In some respects, to be sure, energy and economic costs are inseparably fused. 



1 Atomic Industrial Forum, "Info." March 197.1. 



66 

Rising energy costs have been permeating the system, generating higher eco- 
nomic costs and inflation. This is one reason why many companies recently have 
balked at going ahead with grandiose energy projects they planned — oil shale 
development, coal gasification, nuclear power plant construction and the like. And 
why some pollution control projects are in trouble as well. 

How did we drift into our present energy dilemma? In large part because we 
took for granted an endless supply of cheap fossil fuel energy. As Dr. Howard T. 
Odum, a professor of environmental engineering at the University of Florida, 
has noted : "Only the last two centuries have seen a burst of temporary growth 
made possible by the one-time use of special energy supplies that accumulated 
over long periods of geological time." a 

These bountiful energy supplies offered a means to boost labor productivity 
and profitability. So industries were seduced into squandering energy and shift- 
ing to capital and energy-intensive processes at the expense of labor. 

The United States' highly mechanized agriculture is a conspicuous example. 
(It has been estimated that the agriculture industry consumes more than five 
times the energy content of the food it produces. 3 ) 

So we have locked ourselves into an economy heavily dependent on intensive 
u^e of energy, and we can expect to have considerable trouble making necessary 
adjustments. 

To understand how energy analysis can aid public and private decision-making 
in these days of energy constriction and rocketing costs, it is important to ex- 
umine some basic facts about enegy. In the first place, an energy transaction can 
seldom be measured with the same neat precision as an economic transaction 
expressed in dollars. There are different kinds of energy, different sources, differ- 
ent uses and different ways of measuring. 

For one thing, we have energy stored in structural states — in the chemical 
structure of a lump of coal, in the cellulose of a piece of wood, in the hydrogen 
atom, in the water in a reservoir. And we also have energy in kinetic states — 
Buch as electricity, heat and motion. 

Energy is constantly being transformed from structural to kinetic states (coal 
changed into electricity) or vice versa (sunlight embodied in a plant). So energy 
in a system can be tracked in the form of materials as well as pure energy forms. 

Under the First Law of Thermodynamics, the law of conservation, the total 
amount of energy is constant. But its forms change, and there are differences in 
the amount of work and types of work that can be done by various forms. Oil 
is a highly valued energy species. You can run cars on it. Natural gas is even 
more desirable. It can be easily piped back and forth and can be used for many 
purposes, including electricity generation, home heating and industrial processes. 
Electricity can run computers. 

So there are major differences in the value per Btu of energy forms. And 
under the Second Law of Thermodynamics, the law of entropy, there will always 
be, in any energy transaction or transformation, a net overall loss in the quality 
of the energy (even though part of the output, such as electricity generated from 
coal, may be of higher quality). Entropy can be defined as an increase in the 
Inability to do work. 

There is zero entropy in gravitation. But terrestrial waste heat has lost almost 
all its ability to do work. As changes occur in the universe, energy continually 
slides down the scale, and entropy increases. 4 

The most frequent and well-known use of energy analysis is to calculate the 
total cost of producing and operating something, and then compare it with alter- 
natives. This has been done in analyses of manufacturing, agriculture, 6 trans- 



2 Not Man Apart, mid-August 1974. 

3 The Sciences. New York Academy of Sciences. October 1973. 

4 Entropy is discussed by Nicholas Georgescu-Roegen at length in the magazine Ecologist, 
June 1975'. and at greater length in The Entropy Law and the Economic Process, Har- 
vard University Press. 1971. 

5 Energy Consumption in Manufacturing, by The Conference Board, a report of the 
Ford Foundation's Energy Policy Project, Ballinger. 1974. 

8 For example, "Energv Use in the U.S. Food System," by John S. Steinhart and 
Carol E. Steinhart, Science, April 19, 1974 : "Food Production and the Energy Crisis." 
by David Pimental, et al. Science, November 2. 1973 : "Energy Subsidy as a Criterion 
in Food Policy Planning," by Malcolm Slesser, Journal of the Science of Food and Agri- 
culture, November 1973. 



67 

portation " and energy production itself, 8 as well as in studies of materials/"' ap- 
pliances, 10 bottles (returnable vs. throw-away) 11 and other products. v - 

"In general, a more efficient device will cost more energy to produce," says 
Milton D. Rubin, of the Raytheon Co. "However, usually it will be found that 
the extra energy necessary to produce the more efficient device will be paid for 
very rapidly in the operating savings." 13 

The air conditioner provides a theoretical example : to double the energy effi- 
ciency it would perhaps be necessary to double the weight and thus the energy 
needed for manufacture : but the extra energy cost might be saved by six months 
of operation. (The monetary cost could be another matter. ) But in the case of the 
automobile, as Rubin points out, a smaller, lighter car will not only consume 
less energy in operation, but require much less to manufacture. (And cost less 
as well.) 

Energy analysis has often been used to quantify in energy units the value of 
natural systems. For example, the calories of sunlight that fall on an area, or 
that are captured by plants through photosynthesis. But analysts must beware 
of gross calculations of energy that is not really available or usable, and there 
is some disagreement among them over methodology. Dwain Winters, a program 
analyst with the Environmental Protection Agency, explains a widely accepted 
technique, using solar energy and wind as examples : 

"Although nature provides us with vast quantities of solar energy every- 
day, this does not necessarily make it an abundant energy source. For this 
energy to serve the needs of a technological .society, it must be subject to con- 
centration and storage. Therefore, the measure of solar energy's availability 
lies in the relationship between the energy cost of the equipment used to con- 
centrate and store the energy and the gross energy output of that equipment." 
He adds that analysis now in progress suggests that a significant amount of 
net solar energy probably will be available. 

"For people who dry clothes on a line, the wind is an important source of 
energy," says Winters. "You could figure out the total kinetic energy of the 
wind. But that would not be its energy value to society. What we need to know 
is its energy opportunity cost. That is the amount of energy that would be 
needed to do the same work by a substitute method, such as a clothes dryer. So 
the value of the wind is not fixed, but varies with the energy cost of the tech- 
nology that is used to replace the wind's function. 

"The methodology applies equally well to any energy sources." 

It has been noted that different energies have different values. One indication 
of value is the kind and amount of work the energy can do. Another is its rela- 
tive availability — how abundant is the energy, and how easy to extract. 

This points to one of the problems which must be considered in an energy 
analysis — the "mixed fuels problem." Says Winters : "If we take 1,000 Btu's of 
natural gas and compare it with 1,000 Btu's of oil or coal, we can see that they 
are not of equal value to society. They are not interchangeable in their work 
tasks, nor does their use create equal environmental or social costs. In .some 
cases, adding different energy species together raises serious questions about 
what the results really mean." 

The search for higher quality energy can lead to perfectly rational decisions 
to accept a very small net energy gain or even "negative net energy." For an 
example, Winters takes an imaginary energy analysis of oil shale. ,: Supi>ose it 
showed oil shale to # be a negative net producer. Ignoring for the moment the 
environmental and economic aspects, we still might want to go ahead with 



Conservation and Efficient Use of Energy." joint hearings of House Government 
Operations subcommittee on conservation and natural resources and Science and Wr<> 
nautics subcommittee on energy. Part 2. July 10. 1973: and The Energy Conservation 
Y£?? r *> edIte d hv Robert H. Williams, a report of the Energy Policy Project. Ball i riper . 

m'v'W?. 8 of Energy and the Energy of Systems." by Thomas A. Robertson. Sierra 
Club Bulletin March 1975 : "The Energy Cost of Fuels." by P. F. Chapman et al. Energy 
Policy. September 1974. 



L'w b £ En t r £y Costs of Materials." by P. F. Chapman. Energv Policv. March 197.'. 

1 Hidden Waste.* by David B. Large, The Conservation Foundation" 1<»7:: 
• •» **, l A. CP -" -A bookl , P t of tne I *«*n€ «> f Women Voters Educational Fund. 1975 1 

«U Cans - Energy.' by Bruce M. Ban n on. Environment March 18 
1974 eXample ' " Goods aml Services." by David J. Wright. Energy Policy December 

1975 Amerl0an Association for tne Advancement of Science, annual meeting. January 30. 



68 

oil shale production if society places a high enough value on the resulting liquid 
hydrocarbon fuel. 

"In such a case, oil shale might be particularly attractive if it could be 
subsidized by an abundant energy source such as coal, and if it represented a 
more efficient way to convert coal to synthetic oil than direct coal liquification." 

It should be remembered, however, that while it may be practical to have 
a significant net energy loser in an energy system, the system as a whole must 
deliver a high net yield. 

There can be reasons other than energy quality to accept less net energy. One 
is geographic. Society may justify a heavy expenditure of energy to extract and 
transport fuel to a distant city where it is most needed. Another consideration 
is time. Society could condone a net energy loss to gain time to develop a promis- 
ing new source. Thus the United States can be viewed as encouraging nuclear 
power in the hope that it will tide us over until fusion, solar energy or some 
other technique is readily available. 




From the start of planning to the completion of construction (at line A), an 
energy production plant takes a certain amount of energy from society. At 
some point after the plant has begun operation, it will have paid back an 
amount equal to what it borrowed. That point is at Line B. Thereafter, the 
net energy gain to society depends on how long the plant is kept in operation. 

Much energy analysis, in fact, is focused on the costs of obtaining energy it- 
self. One study concluded that in the United Kingdom, five energy industries 
(coal mining, oil refining, coke, gas and electricity generation) jointly consume 
more than 30% of the UK's total energy input. 14 

Of course the problem is exacerbated in the U.S. as reserves such as oil and 
gas become increasingly unyielding and costly to extract (not the case with 
coal stripped from land in the West, to be sure), and as more primary fuels 
are converted to secondary energy sources (gasification of coal, for example). 
"Many proposed alternative energy sources would take even more energy feed- 
back than is required in present processes," says Odum. 

Winters points out that if we are planning on shifting to a solar or nuclear 
economy, "we need to know to what extent such an economy is self-sustaining 
and where it is dependent on fossil fuels. If we're going to make these tran- 



14 "The Energy Cost of Fuels,' 
1974. 



by P. F. Chapman et al, Energy Policy, September 



69 

sitions, we want to make sure we've made the appropriate use of our fossil 
fuels to get us from here to there." 

Winters notes that during the planning and construction of any energy pro- 
duction facility, the project is borrowing energy from society. "When the plant 
begins operation it starts paying back this debt. If the plant is a net energy 
producer, eventually it will pay back this debt and then contribute a net energy 
gain to society until the plant wears out. For any energy process, there is this 
period during which it is a net energy sink. Just how long the period lasts 
should be of considerable interest." 

The chart on page 3 illustrates a theoretical energy sink, payback period, and 
net gain in energy. "We need to know the payback period relative to different 
energy species," says Winters. 

The length should be of particular interest with regard to nuclear power tech- 
nologies. For even if a nuclear plant should turn out to be a substantial net 
energy gainer, it could still create a net energy problem. If nuclear plants are 
built in too rapid succession, there might be temporary energy shortfalls because 
of the immediate demand for energy to build the plants. 

A workshop of the International Federation of Institutes for Advanced Study 
(IFIAS) offers this example (with the figures picked for illustrative purposes) : 

"Consider a country embarked on a program of nuclear reactors. Suppose 
such reactors to have a 30-year life, and to have a capital energy requirement 
to build them equal to 10% of the total energy production in their life ; that it 
takes six years to build a reactor, and one new reactor is started each year . . . 
For many years the country will incur a net energy deficit (11 years), and 20 
years will elapse before the cumulative energy production exceeds the cumula- 
tive energy investment." 15 

Two parallel debates over nuclear power are being waged — one over its 
energy efficiency and one over its cost efficiency. Davis, of Bechtel Power, esti- 
mates that all the energy invested in a nuclear plant is repaid after 2.3 months 
of full power operation. 1 Similarly, many utility executives have touted nukes 
as great cost savers over coal-fired plants. 

Nuclear critics, on the other hand, charge that industry calculations are care- 
fully tailored, and don't include all the money and energy costs (or subsidies) 
associated with mining uranium ore. enriching nuclear fuel, reprocessing, re 
search and development, insurance, safety, plant decommissioning, and long- 
term waste disposal and safeguarding. 

"When all energy inputs are considered," says energy consultant E. J. Hoffman, 
"the net electrical yield from fission is very low." 16 

Much of the difference l>etween the various estimates lies in what the analysts 
include in their accounting and what they leave out. In other words, where they 
draw their system boundaries. And that is a critical decision in any energy 
analysis. "If you let me draw the system boundaries wherever I want. I can 
make almost any project look good or bad." says Winters. "That's the nature of 
the game." 

He says the boundary is determined in part by the question you're trying 
to ask. "You may be interested only in the energy you're borrowing from society 
as opposed to the amount you're returning. You may be interested in the rate at 
which you're depleting a resource ... In that case, what you do is put the 
resource that's in the ground within the system boundary of society." 

Or, in accounting for the energy cost of an automobile, do you include the gas 
an assembly line worker use* getting to the job? The energy used to make his 
automobile? The gas used by those who made it? Perhaps the most important 
thing an analyst can do is make very clear what his boundaries are. 

Energy analysis can help identify the most fruitful conservation strategies. 
It can compare the savings from home insulation with the energy costs of mak- 
ing and installing the insulation. It can pinpoint ways to reduce the energy in- 
tensiveness of a manufacturing process and therebv lower the energy input per 
unit. 

Winters says that an overall conservation strategy is suggested by the fact 
that some 70% of all energy used by a final consumer is in the form of 'goods and 
services, rather than species of energy such as electricity and gasoline. "So 
if we really want to cut down our consumption, there's probably more room 

25-Vo E % r 74 Anal - vsis -" Workshop Report No. (5, IFIAS, Guldsmedshyttan. Sweden, August 

»• Quoted in "It Takes Enerpy to Get Energy," by Wilson Clark. Smithsonian. De- 
cember 1974. 



70 

to attack that sector than there is in how much gasoline we use or how much we 
heat our houses. What this implies is somehow changing the product mix." 
Such as shifting to returnable bottles or smaller cars. 

Energy analysis, says the IFIAS workshop report, is a means of identifying 
the constraints of a system. "For example, by means of thermodynamic calcula- 
tions one may establish the theoretical energy requirements for a process, and 
compare them with those of present technology. This gives one a feel for the extent 
to which a given technology could be developed — a limit not identifiable by 
economic methods." 

Energy analysts certainly don't see their techniques as a replacement for 
economics — though the limitations and failures in trying to apply economic 
analysis and operate a market economy have been all too obvious. (One thinks 
of the widespread anti-competitive practices and arrangements, including OPEC ; 
the natural monopolies like public utilities ; the government regulation induced 
by varied corporate abuses ; the distorted flow of information ; the failures to 
account for external effects such as pollution ; and the many subsidies provided 
by taxpayers. The effects include social inequities, environmental degradation, 
resource scarcities, recession and inflation.) 

Winters suggests that an economic analysis can be made of an individual 
technology — or a mix of technologies — to check feasibility, and this can be 
followed by an energy analysis to ascertain efficiency. 

The IFIAS group is particularly interested in the use of energy analysis to 
understand price changes and other factors in the economic system. It can 
be a "means of injecting physical variables into economic theory," the workshop 
report said. "It can be a more sensitive indicator than money." 

The IFIAS group is planning a follow-up conference "to consider the interface 
between energy analysis and economics." 

Winters sees at least five "gates" through which a decision-making process 
must pass to show that a project will work : political, economic, cultural, envi- 
ronmental and energetic. 

"The two primary parameters that are currently being used are what is polit- 
ically feasible and what is economically feasible," he says. "They don't always 
overlap. These are the two areas in which most of the arguments over weeding out 
the various energy alternatives takes place." But the constraints of the other 
three gates must be satisfied in order to conform to human behavior, to main- 
tain the stability of the environment, and to be thermodynamically, or physically, 
possible. 

There is reason to believe that the energy gate is one of the narrowest of all 
the gates for society's options to pass through. So if limited funds are available 
to evaluate options, it is logical to ascertain early what will fall through the 
energy gate. 

Winters notes that perhaps environmental considerations should sometimes 
be grounds for rejecting energy options, but may not carry the necessary political 
clout. "Whereas, if one took those environmental objections, and at the same 
time noted that something is energetically impractical, one might be better able 
to implement environmental criteria." 

One of the more interesting questions raised by net energy analysis is what 
are the cybernetic (feedback control) properties of an energy system. If it takes 
energy to make energy, to what extent do these energy subsidies control those 
functions outside the control of economic forces? 

In the past, such influences may have been minor, but in a period of resource 
scarcity, their influence could be dominant. To illustrate how we might better 
understand and manipulate to our advantage these cybernetic functions, Winters 
uses a hypothetical ecosystem as an example. (An ecosystem is a highly complex 
energy system which is governed by its energy cybernetics.) 

"You can think of any animal as an energy storage. Take for example a group 
of lions in a closed ecosystem. They are in a highly ordered state, tending to go 
toward disorder — an example of the law of entropy at work. And so they must 
have an energy source to maintain their order. Otherwise you have fewer lions, 
or emaciated lions. 

INDUSTRIES ARE JUDGED BY ENERGY THEY USE AND JOBS THEY CREATE 

These days particularly, there is good reason to factor into energy analysis 
the effects on employment. Because typically, reductions in energy use generate 
more jobs. John P. Holdren, assistant professor at the University of California, 



71 

Berkeley, says that the energy producing industries themselves comprise the most 
capital-intensive and least labor-intensive major sector of the economy. 

"Accordingly," he says, "each dollar of investment capital taken out of 
energy production and invested in something else, and each personal-consumption 
dollar saved by reduced energy use and spent elsewhere in the economy, will 
create more jobs than are lost." 15 (Whether that dollar helps or hurts the energy 
situation would depend on where it's spent.) 

Much valuable work in this area has been done by Bruce Hannon, of the Center 
for Advanced Computation, University of Illinois at Urbana-Champaign, One case 
study showed that returnable bottles demanded less energy and more labor 
than throwaways. 17 

But Hannon has gone much further. He has plotted on graphs the relationships 
in 363 industries between dollar value added, energy use and employment. 18 

One chart shows the tradeoffs in the economy as a whole resulting from a 
10% growth in specific industries, if total GNP is held constant. For example, 
it indicates that growth in railroad car manufacture would have an overall 
impact of increasing energy use by almost 4.5 trillion Btu's and decreasing employ- 
ment by more than 2.5 million jobs. Toward the other end of the scale, a 10% 
growth in making wood furniture would lower energy demand by more than 
a trillion Btu's while increasing jobs by more than 2.5 million. 

Such analyses suggest that it is possible to make intelligent plans to shift from 
enterprises that are negative in terms of energy consumption, environmental 
effects and job formation, to enterprises that better serve society's needs. 

In some cases, existing industries, by eliminating waste and altering products, 
could make desirable transitions. In most cases, perhaps, change would be ex- 
tremely difficult. Ecologist and author Barry Commoner has long stressed the need 
for shifting away from products — such as plastics — that involve heavy pollution 
and energy consumption. But he notes that with some industries, such as the 
petrochemical makers, "the intensive use of energy is built into the very design 
of the enterprise in order to eliminate human labor." 18 

Even if a large net gain in employment is indicated, the type of labor or its 
location could be radically different and pose an additional dilemma. Of the 
shift to returnable bottles, Hannon says : "Jobs would be lost in the highly 
organized, high-wage can makers' plant and gained in the low-wage, relatively 
non-organized retail sector." Thus the opposition to the plan by organized 
labor. 20 

"That's something that labor will have to deal with," says Nick Apostola, 
coordinator of a new organization, Environmentalists For Full Employment 
(EFFE), located in Washington, D.C. He suggests that labor "seize the oppor- 
tunity to boost up those low-level jobs so they'll be higher paid." In a recent 
policy statement, EFFE said : 

"Modern technologies that are excessively capital intensive and energy waste- 
ful simultaneously destroy the environment, deplete resources, and cause struc- 
tural unemployment. These problems must be attacked concurrently and such 
technologies must be rejected." 

"Suppose they pick zebras for an energy source. In order to utilize the zebra 
energy — since zebras don't just walk into the lion's mouths — the lions are going 
to have to expend some energy to run down the zebras. I used to be a zoo- 
keeper and have some appreciation for the task the lion has before him. 

"If our lions are to survive, they must make sure they get more energy out of 
zebra hunting than they put in. Let's say they put two energy units in and get 
10 out. 

"A lion will be competing with some other carnivore in that ecosystem for an 
energy source. Maybe the leopard. And one factor that determines who will win 
the competition is who has the most efficient return on his investment. If another 
carnivore in competition with the lion could reinvest at a rate of one unit in for 
10 out, then he would end up being the dominant sj>ecies. 

"In the ecosystem, no two animals occupy the same niche, competing for the 
same energy source. There's always at least a slight difference. And the relation- 
ship becomes very complex. Finding the niche with the best reinvestment ratio 
becomes a major key to species survival. 



" New York Times. July 23. 1975. 

18 "Options for Energy Conservation." reprinted in Technology Review. February 1974. 
18 "The Energy Crisis — All of a Piece." Center Magazine. March-April 1975. 
20 "Energy Conservation and the Consumer," Center for Advanced Computation, Unlver 
sity of Illinois, October 1974. 



72 

"It carries with it a risk, however, for if our lions should become solely depend- 
ent on zebras, then what happens when there is a zebra plague? The answer, of 
course, is that the lions eat something else. To be able to compete for other prey, 
however, means that the lions have had to maintain an ability to hunt for ani- 
mals which did not yield a maximum return in their investment. And if we 
examine lion habits, we see that they have done exactly that. 

"This makes the problem most interesting, for it implies that there is a force 
in addition to net energy efficiency directing lion evolution. This force is called 
diversity. Diversity, in the case of lions, is seen in their varied feeding habits. 
It enables them to survive environmental changes. So we see two antagonistic 
forces operating upon lions : one, hunting efficiency strengthening their chance 
for survival in the present and two, diversity enhancing chances of survival in 
the future. 

"A given organism has to find a compromise between how much energy it's 
going to tie up in reinvestment and how much in diversity. If it ties up in rein- 
vestment it may become a dominant member of the ecosystem at any given time, 
but may also end up the way of the dinosaur. 

"The application to man's technological system, I think, is direct. We should 
choose energy systems which show a good net energy gain, but we must balance 
this with a diversity of interchangeable technology which will help protect us 
from the unforeseen problems that always interrupt the best-laid plans of men 
and lions." 

Winters points out that man, throughout history, has periodically improved 
his reinvestment ratio. He learned to use rocks to break bones and get the mar- 
row. And there was agriculture and fire, and so forth, up through the stream of 
innovations in the Industrial Revolution. But after each of the earlier innova- 
tions there appears to have been a long plateau, so that energy crises were more 
the rule than the exception. 

Where are we now? At one of the "little wrinkles" of a major growth spurt 
that will continue for a long time? Or are we approaching a leveling spot of one 
of the broad plateaus, in a situation where we have hegun to exceed the total 
productive capacity of our energy systems? 

Winters says that to maintain our growth rate in energy supply, the techno- 
logical innovations must be bigger and come in more quickly than in the past. 
Energy analysis may tell us the theoretical upper limits of available energy, but 
physical limits apparently are not the prime constraint. "If we look at today," 
says Winters, "we know we're not at our physical limits in terms of what we 
could do theoretically. It looks as if social and cultural limits always begin to 
impede the system before the technical limits are reached. 

"The question may not at all be, Can we build a fast-breeder reactor? Or can 
we get nuclear fusion? It may well be, Can we construct social institutions which 
can maintain the dynamic stability of the cultural, political and economic proc- 
esses that are needed to maintain such a technology ?" 

As a matter of fact, Winters suggests that the social structure we have come 
to depend on — with its standard of living and complex, centralized system of 
organization — may itself demand a lot of energy to maintain. 

"There is an interesting paradox — and this is pure supposition and beyond the 
level of measurement. If we believe there is some relationship within a given 
system between the complexity of a social order and the energy base (and one 
culture may require more energy than another), then in order to get total cooper- 
ation at any larger level of complexity, a broader energy base is required." An 
example would be the need to achieve affluence t>efore people cooperate in popu- 
lation control. 

"If it takes affluence to induce cooperation, and you can only do this by increas- 
ing the energy base (a necessary but not sufficient cause for affluence), then it 
would seem that the only way to increase the energy base is through greater 
cooperation. So you've just got yourself in a self -stultifying situation." 



APPENDIX II 

Critics Begin to Surface 

Following several years of development and promise, the limitations 
of energy accounting have begun to emerge. The three articles which 
follow ask what the technique offers that economic analysis fails to 
deliver and conclude that, given the present state of development, it 
not only offers little but might even be misleading. 



Energy Accounting vs. the Market 

Based on a paper by Joel Darmstadter, given at the Pacific Science 
Congress, Vancouver, B.C., August 22, 1915. The paper was a spinoff from 
Darmstadter's recently published RFF study, Conserving Energy : Pros- 
pects and Opportunities in the New York Region (see p. 7) . 

The concept of energy accounting, which is now enjoying something of a vogue, 
challenges the adequacy of the usual economic forces and economic criteria to 
impel desirable energy conservation measures. Its proponents argue that rational 
energy conservation measures can only be pursued in the framework of explicit 
and quantitatively detailed knowledge of the energy implications of contemplated 
courses of action. In this view, for example, the choice between a bus system 
and a rapid-rail transit would be decisively governed by which mode promises 
to deliver the most passenger miles per unit of energy required. Any other cost 
questions would be secondary. In its extreme form, such an approach would 
accord to an "energy standard of value" a status parallel to the conventional 
monetary basis of decision making. Indeed, some legislators have sought to impose 
a requirement for energy impact statements on contemplated governmental 
actions. 

One of the concepts spawned by energy accounting is that of "net energy." 
which refers to the fact that it takes energy to produce energy. While this is 
obviously true, it is less obvious that energy input-output measures are the only 
way to avoid the dangers of putting more energy in than is gotten out. 

The dissatisfaction the proponents of energy accounting feel toward market 
forces has arisen for obvious reasons. It is easy to observe that people often 
prefer to drive their cars to work, even when public transportation is cheaper 
and readily available. Manufacturers may be reluctant to switch from an exist- 
ing production process to another, more expensive but less energy-wasteful, mode. 
People with frost-free refrigerators are willing to pay not only for the more 
expensive appliance, but the higher electricity bills as well. Clearly neither cost 
saving nor energy conservation ideals are the prime considerations of the com- 
muting car drivers and the frost-free refrigerator owners. And the manufacturer 
in the example is actually rewarded for his energy-intensiveness. 

It would be incorrect, however, to assume that these observable behavior pat- 
terns are an argument against using cost and price measures for energy con- 
servation. Rather, they are an argument for remedying a defect in market pricing, 
which has failed to exact full payment for social harm. For example, no pay- 
ment has been exacted for environmental damage caused by automobiles. A manu- 
facturer's energy use might so pollute the environment that, were he charged for 
the damaging emissions, his inducement to shift away from energy-intensive 
processes might rise. Those using energy-wasting appliances have chosen to pay 
higher energy costs to avoid drudgery — but here, too, costs might, in time, rise 
to a point that would discourage all but the most luxury-loving. In the last 
analysis, choices to conserve energy can be influenced by higher costs, but cannot 
b a dictated by them because energy consumption is only one element in a much 
wider range of human activity. Consequently, the energy accounting approach, 
by focusing solely on energy, may result in misguided policy emphasis. 

An example of this can be shown in land use policy. Cities, with their high-rise 
apartments, compact shopping districts, and public transport, are much more 
energy-efficient than suburbs, with their heavy use of private auto transport, 
dispersed shopping facilities, and single-family dwellings that are considerably 
heavier energy users per square foot of living space than multifamily units. But. 
clearly, energy consumption is only one of many social concerns in land use : be- 
fore choosing to increase the number and density of cities, one would wish to 
consider such things as esthetics, recreation, open space, conserving resources 
other than energy, less air pollution, less crime, and so forth. The costs of these 
amenities include some degree of increased energy use, but the advantage of the 
market pricing system, when it works properly, is that it takes all costs — not just 
energy — into account. 

(75) 



76 

This does not deny, however, the existence of imperfections in market pricing 
processes and institutions, as well as gaps in information, which do interfere 
substantially with desirable shifts toward energy-conserving behavior. To rem- 
edy these defects, a number of measures to improve the market pricing approach 
could be considered. 

For energy users to be able to respond knowledgeably to market conditions, 
government policies designed to guide consumption practices along a more in- 
formed path are clearly desirable. Examples include mandatory information on 
energy efficiency and costs in the heating and cooling of newly constructed build- 
ings ; in the operation of automobiles ; or in the use of room air conditioners. 

Governmental policies designed to shape market outcomes through explicit 
action to influence prices could also be brought into play. For example, a govern- 
mental horsepower or weight tax would help sway owners towards smaller cars. 
Gasoline taxes would do the same and might encourage much more car pooling. 
The expansion of public transport— particularly bus transit systems, which are 
less burdensome than is the case with the enormous capital commitment of rapid- 
rail service — is badly needed. In housing, compulsory insulation standards and — 
conceivably — some changes in home financing arrangements supportive of energy 
conservation practices suggest themselves. 

In the past, governmental action may have had an effect detrimental to con- 
servation. For example, many people believe that the government's control over 
interstate natural gas prices — which are set far below market clearing levels — 
have artificially encouraged consumption of the scarce and desirable resource 
and deterred supply expansion. 

In J summary, the potential for significantly diminished levels and growth rates 
in energy demand undoubtedly exist, but the feasibility of such savings must 
be measured by an economic yardstick. The virtues of the price system, with all 
its weaknesses, is that all costs — not just energy costs — are accounted for. Only 
through the medium of the price system can we measure the benefits gained or 
foregone by altering a consumption habit or a production process involving direct 
or indirect energy inputs, rising energy costs can and probably will curb demand 
(there is also the possibility that some part of the public will meet rising en- 
ergy costs by curtailing expenditures in other areas). Public policy can do much 
more in fostering conservation through information and demonstration pro- 
grams and through tax and price measures that expose the energy consumer to the 
full cost of his consumption. 



The Economics of Energy Analysis 
(By Michael Webb and David Pearce) 

The object of this paper is a critical appraisal, from the economist's standpoint, 
of whjat appears now to be called energy analysis, but which has, at one time or 
another, been called energy budgeting, energy accounting and energy costs. The 
proliferation of articles on energy analysis and the fact that it has apparently 
been afforded serious attention in political circles and official documents, 1 are in- 
dications of the importance now attached to this technique. Remarkably, how- 
ever, eneigy analysis has been subjected to only a minute amount of published 
criticism, although we are aware of extensive verbal criticism from many quar- 
ters. We seek to correct the balance as far as the published literature is concerned. 
While what we have to say is frequently critical in a purely negative sense, we 
hope that what we have to say will prompt energy analysts to define in a rigorous 
fashion the real uses of their studies. 

THE GENESIS OF ENERGY ANALYSIS 

It is useful to begin by asking why energy analysis should, in recent years, 
have become so important, even if that importance, in our view, is exaggerated. 
The genesis of EA is clearly the awareness, now a commonplace, that the world 
has many natural resources in finite supply. Of these, energy resources — at least 
as far as the fossil fuels and uranium are concerned — are clearly limited as a 
stock. Others such as solar energy are limitless flows (within any sane time hori- 



1 National Economic Development Office, Energy conservation in the United Kingdom: 
Achievements, Aims and Options, HMSO. London 1974. See especially pp 90-1. 



77 

zon anyway). However, as far as the 'renewable' or unlimited fuels are concerned, 
technology is not developed to apply them to practical use and they must remain 
in the realm of probabilistic future supplies. To plan the future on the basis of 
what might be is to engage in a maximax policy the costs of which, if technology 
fails to generate the required new fuels, would be catastrophic to future 
generations. 

We accept that the critical issue in what we might call nonspeculative plan- 
ning is that of rationing resources intertemporallly in the most equitable fashion. 
Inter-generational equity is a subject of detailed concern by economists, al- 
though we would be dishonest if we suggested that all economists believe in ac- 
commodating the problems bequeathed by one generation to another in their 
analysis. But what we can say is that energy analysis offers no assistance with the 
problems of intertemporally allocating resources in finite supply. 

The reason for this is simple — EA is a purely mechanistic technique devoid 
of all value content (even though, as we shall argue, its exponents often use it 
for evaluative purposes). As such it can tell us nothing about optimal allocation 
rules. We make no specific counter-claim for economic analysis either. The liter- 
ature on the economics of intertemporal allocation is frequently rarified and this 
is not the place to discuss its merits. 2 What we are concerned to note is that the 
fact, if it be one, that current technoloy sets a limit to the availability of exploit- 
able energy is not one that affords EA any importance. Instead, it seems to us 
to make the normative issue of intertemporal allocation that much more impor- 
tant, and this is not something that EA can, in our view, assist. 

Slesser 3 writes that 'economics treats the world as a closed system having ac- 
cess to limitless amounts of energy, whose acquisition takes only time, capital, 
labour and technology. But this is simply false. There is nothing in economic sci- 
ence that requires us to assume limitless resources of any kind. Scarcity is, in fact, 
the very foundation of economics. 

The same false charge that economics assumes 'no shortages of any inputs 
to the production system' has been made by Chapman. 4 He compounds the error 
by stating that economics assumes substitutability between inputs wiiere as en- 
ergy analysis does not. Indeed, both Slesser and Chapman emphasise that, if capi- 
tal and material inputs are reduced to energy units, the result is a two-input 
model of the economic system — the two inputs being labour and energy — then 
there will be non-substitutability between labour and energy. As Slesser says 'in 
the last analysis, energy does what labour cannot do". Continuous substitutability 
between input is indeed an assumuption of neoclassical economic analysis, but it is 
not a necessary one and much of the progress in economics since the formulation of 
neoclassical axioms from 1870 to 1940 has been in the realm of reassessing the 
neoclassical lesults in the context of product discontinuity. The general outcome 
of this analysis is that the original results of the neoclassical system remain 
intact. 

However, even if the possibilities of substitution are non-existent, it remains 
the case that the problem again reduces to one of rationing resources over time. 
Thus, if labour and energy are, in some sense, the only resources, if substitution 
is constrained, and if energy is a finite non-renewable stock whereas labour is a 
renewable flow of resource, the policy options are : 

Allocate resources over time according to some intergenerational welfare 
criterion. 

And/or restructure the configuration of material output in favour of 
labour-intensive and against energy-intensive outputs. 

It is evident, and would seem to be admitted by most energy analysts, that 
EA has nothing to tell us as far as the first option is concerned. Its role in the 
second also seems more than questionable. If energy is scarce then some price 
change will take place, generating exactly the product substitution called for. 
However, if, as we would accept, market prices are non-optimal even with re- 
spect to current-generation biassed decision-rules (and hence even more non- 
optimal with respect to future-oriented rules) then it would be useful to identify 
energy intensive activities so as to adjust market prices to reflect the true 
shadow price, perhaps, by the use of an energy tax. 



2 For an interesting contrast sop the essays bv I. F. Pearee and by J. Kay and J. 
Mirrlees in D.W. Pearee (ed). The Economics of Natural Resource Depletion, (Maemillan, 
Basingstoke. 1975). 

3 M. Slesser. ' Accounting for energy'. Nature, vol 2. r )4. March 20, 197f). 

* P. Chapman, 'Energy costs : a review of methods'. Energy Policy, No. 2. vol 2, p 91 



78 

What is not clear, however, is why we need EA to identify such energy- 
intensive uses : a tax on energy consumption can be implemented without carrying 
out elaborate exercises to identify energy use. If, say, some tax proportionate 
to energy consumption was introduced, energy-intensive activities would auto- 
matically bear the heaviest tax burden, simply because energy costs comprise 
part of the costs of production of economic activity and because these costs are 
shifted forward from the most basic economic sectors such as resource extraction 
to the final product. In short, we fail to see where the 'two factor' approach 
adopted by energy analysts takes us. At best it adds nothing to what simple eco- 
nomic analysis tells us, at worst it serves only to obscure the issue. 

We may also note at this point that reducing inputs to energy and labour ob- 
scures the fundamental reason for the separating out of capital by economists. 
Basically, capital generates a flow of goods in excess of the original value of the 
capital. Indeed, this is the very rationale of capital investment. If it were not. 
we would merely be diverting resources from one use to another with the same 
value and hence gaining nothing. We have nowhere seen in the EA literature ref- 
erence to the productivity of capital. 

Clearly it can be accommodated in the sense that capital reduced to energy 
units generates a flow of energy values if the project is productive of energy 
(eg a nuclear power station). We would then have a situation in which there 
would be labour inputs (measured in man-hours or in value terms?) plus energy 
inputs to be offset against energy outputs plus any non-energy outputs. But such 
a calculus is devoid of any use, because it offers us no decision rule by which to 
choose options. If we have a project using a combination of, say, 4 energy units 
and 6 labour units, with output 5 energy units and 3 units of some commodity, 
how do we decide if it is worth undertaking? 

The absence of a homogeneous measuring rod relating to values in energy 
analysis makes such a calculation worthless. The only way homogeneity of units 
could be achieved would be to reduce labour to energy terms as well and to value 
commodity output in energy terms. The latter appears to be what some energy 
analysis would want to do, but they resist the reduction of labour to energy units. 
As such we have no homogeneous unit and hence no decision criterion. 

Alternatively, the rule might be to select projects which minimise energy costs 
regardless of the levels of other inputs. Energy analysts appear divided on 
whether such a rule is sensible. While protesting on the one hand that EA 
cannot be evaluative it is not in the least difficult to find in the literature state- 
ments such as 'energy analysts believe that it makes sense to measure the cost 
of things done, not in money, which is after all nothing more than a highly 
sophisticated value judgement, but in terms of thermodynamic potential.' 5 But 
if EA is not evaluative, what is the point of such an elaborate exercise? 

We elaborate on this point in later sections. For the moment we argue that 
the energy analyst's attempt to differentiate their subject from others by reducing 
economic systems to two-input analysis serves no useful function and, indeed, 
only obscures some important aspects that differentiate energy embodied in 
capital from energy embodied in other commodities. 

There may well be other 'energy limits' or 'boundaries' to economic activity. 
Many writers have commented on the pollution limits set by the effects of waste 
heat dissipation. Georgescu-Roegen for example states 'The additional heat into 
which all energy of terrestrial origin is ultimately transformed when used by 
man is apt to upset the delicate thermodynamic balance of the globe. 5 These 
expressions of concern are well-taken and do indeed reflect lack of attention 
to the laws of thermodynamics by economists. What is not clear, however, is why 
we need EA to identify this boundary, or, if we do need it, why the existence of 
such a boundary is thought to involve some deep criticism of economic analysis. 

The most basic assumption of economics is that economic agents — consumers 
and producers — seek to optimise subject to constraints. If there is indeed a dissi- 
pated heat limit to economic activity this constraint can be added to the optimisa- 
tion problem. Occam's razor demands that we do not multiply our techniques 
beyond the minimum necessary if existing techniques are quite capable of accom- 
modating the problem. 



B N. Georjreseu-Roegen. 'Energy and economic myths'. Southern Economic Journal, vol 
41. No 3 January 1975. 



79 

THE HOMOGENEITY ASSUMPTION AND RELATIVE PRICES 

Energy analysts treat energy as an entity that can be aggregated regardless of 
its source. The exception is labour energy which is differentiated from other 
energy inputs. 

In energy analysis, any commodity, i, can be 'reduced' to some energy content 
Ei plus some labour content Lti. The volume of output can be left in physical, 
monetary or energy terms, giving us expressions of the 'cost' of a unit of that 
output of the form 

Energy Cost ^ ' 

where Qt is the output measure of ?'. We have already commented on the dif- 
ficulties of finding a use for such a ratio. We may note that the aggregation 
problem in economics is 'solved' by using prices. That is, Q? would be measured 
in value terms, and the denominator would appear as 

e = n Lj=Tn 

e=l L=l 

where P e , X e refers to the price and quantity of different energy sources, and 
correspondingly for labour. The prices in these expressions will, if markets 
function properly, reflect consumers' willingness to pay for the product in ques- 
tion, or, for inputs, the benefit foregone to consumers by using the input in its 
current use rather than the next best use (the input's 'opportunity cost'). Where 
markets do not operate properly 6 the prices used are shadow prices — prices 
which, if they did operate, would reflect marginal willingness to pay on the part 
of consumers. Either way, the use of prices serves to homogenise the heterogenous 
units and to import value-content to the resulting aggregate. 

Before proceeding to discuss the validity of the analogous procedure in energy 
analysis — the use of energy units to homogenise inputs — we may note that 
shadow prices also bear some relation to the finitude of resources. Essentially, if 
markets operated freely and perfectly, the current price of a resource would 
reflect expectations about the limited stock of that resource. As the stock is 
depleted the price will rise, thus rationing the use of the resource, inducing the 
adoption of substitutes, encouraging recycling, and so on. Where the totality 
of resources of a specific kind, such as energy, are concerned, this mechanism will 
not operate except in the sense that labour and capital can be substituted for 
energy. 

Now, energy analysts are quite right to point out that the possibilities of this 
kind of substitution are limited, although not. one suspects, as limited as is often 
suggested. We know, however, that markets in natural resources such as energy 
do not function perfectly. We need not dwell on the reasons for this here 7 
but it is important to note that we do not know the extent of the deviation of the 
appropriate shadow prices from the actual market prices. We have noted a 
tendency in some of the EA literature to assert that market prices fail totally 
to reflect future scarcity, an assertion that is nowhere substantiated by any 
evidence. Indeed, assessing the evidence is a complex issue. 

We are prepared to believe, however, that current resource prices are not 
accurately related to future scarcity. What we need to know, then, is how EA 
will assist us in identifying the future limits. As we argued earlier it is not in 
the least clear how it assists. 

We can now turn to the central matter of this section : the homogeneity assump- 
tion in energy analysis. To demonstrate the problems of using a common energy 
unit (or, indeed, any common physical unit) we consider an example which 
contrasts such a physical measure with the economist's concept of opportunity 
cost outlined above. Assume the existence of some resource, call it 'oil', which is 
homogeneous and in finite supply. In addition assume that all the deposits of 



8 'Properly' in this context relates to a situation in which the configuration of prices 
maximises consumers' welfare in the aergregate. The EA lterature contains numerous 
comments on the hiases imparted by using market prices hut we have nowhere noticed 
even on awareness of the idea of shadow pricing, perhaps hecause energy analysts 
mMnkpnly identify economics with the free enterprise ethic. 

7 See the introduction to D.W. Pearce fed). The economics of natural resource deple- 
tion (Macmillan. 1975). where the various divergencies are listed. 



80 

this resource are equally accessible and thus that there is no change in the 
physical inputs required to obtain a ton of this resource. Thus each ton can be 
obtained at a constant expenditure of energy. Finally, for simplicity, assume that 
there are no substitutes available for this resource. 

As the physical exhaustion of this resource approaches, energy analysis will 
continue to measure its cost calculations at a constant 'cost' (in terms or kWht 
etc). But economic analysis would show the price of 'oil' increasing to reflect its 
increasing scarcity. This would probably happen in two ways. Suppliers, seeing 
the coming exhaustion of their product and knowing of the absence of the 
possibility for substitution, will raise its price. Second, if there exists a futures 
market, dealers in this market would offer higher prices for the product as the 
time of its physical exhaustion approached. If there is no futures market, the 
price will nonetheless rise because of supplier reaction. Whether the time-profile 
of prices that results is an 'optimal' profile from the point of view of social welfare 
is not relevant to this example. We noted above that the profile may well deviate 
from such an optimal path. Our point is that prices will rise by some amount to 
reflect scarcity, whereas the energy cost will be constant. 

In such a situation energy analysis would be a poor guide to increasing scar- 
city : indeed, in this case it would indicate no scarcity at all. Now, energy anal- 
ysts have emphasised that one of the main functions of EA is to identify 
changes in relative prices over time. It is argued that, because of the deficiencies 
of the market mechanism, economic analysis will not identify those changes, or, if 
it does, will do so later than energy analysis. 8 Indeed, this appears to be the 
sense of some of the more grandiose claims for EA. Berry, for example, has 
said 'if economists in the market place were to determine their shortages by 
looking further and further into the future, these estimates would come closer and 
closer to the estimates made by their colleagues, the thermodynamicists.' 9 Our 
simple example shows that no such convergent process need occur. 

Now suppose that instead of the 'oil' being equally accessible it has a decreas- 
ing quality gradient — to obtain the marginal barrel of oil we need to expend 
extra amounts of energy and other resources. This is perhaps more pertinent to 
many material resources. As more and more marginal resources are exploited 
we can expect the energy cost to rise and hence there is some relationship be- 
tween scarcity and energy cost. Equally, however, we will find that the real 
economic cost of extracting the marginal resource will rise. We can illustrate 
this by taking the example of copper extraction. 

We know that the grade of copper ore has been declining. It has been calcu- 
lated that fuel input per unit of copper output for the USA fell to about 1930, 
rose slightly from 1930 to 1950 and then rose very fast indeed to 1960. The curve 
is in fact a flat-bottomed 'U' shaped curve. The important question relates to 
the information provided by this curve. Would it indicate that energy analysis 
has pinpointed a rise in the relative (real) price of copper earlier than economic 
analysis would, and hence is more sensitive to scarcity? It is certainly the case 
that if we look at the real price of copper (the money price deflated by an index 
of manufacturing wage rates) we find it rises much later than 1930 — somewhere 
in the mid 1960's. 10 Has energy analysis therefor anticipated the relative price 
rise? 

The problem is that it is impossible to draw any conclusion at all from such 
an analysis. Firstly, the energy cost of copper extraction is only one of the costs 
involved. The energy approach and the economic approach are therefore non- 
comparable. Secondly, if we took the money costs of fuel inputs we would secure 
the same result as EA. Quite simply, whatever the energy inputs into copper ex- 
traction, and however far back the energy costs are traced through the economic 
system, those inputs will have prices and the accumulation of prices will be 
revealed in the final money cost of fuel for copper extraction. In short, if we 
are interested in energy inputs along it is completely unclear why we require 
energy analysis rather than the straightforward and readily obtainable money 
cost of energy inputs. 



8 P. Chapman, 'Energy analysis : A review of methods and applications', Omega forth- 
coming. 

9 R.S. Berry : US Congressional Record. 92nd Congress, S 2430, 1972. Quoted In 
M. Slesser, 'Energy analysis in technology assessment'. Technology Assessment, vol 2, 
No 3, 1974. 

10 W. Nordhaus, 'Resources as a constraint on growth', American Economic Review, 
1974. 



81 

Thirdly, the analysis presumes scarcity rather than demonstrates it. That is, 
the fact that the energy costs of securing extra copper are rising need in no way 
he correlated with a scarcity situation. We argued above that a decreasing qual- 
ity gradient situation would tend to have rising energy costs associated with it. 
It does not follow from this, however, that every situation in which we observe 
rising energy costs is a situation of scarcity. In contrast, whatever the defects 
of market prices, if we observe the real price of copper rising we can deduce 
something about scarcity. 

We must he careful not to claim too much for economic analysis in this respect, 
however, for it is equally true that if markets fail to operate at all sensitively, 
constant or falling real costs might exist even though a scarcity situation might 
exist. All we are saying here is that we fail to see how energy analysis improves 
our knowledge of the situation. Fourthly, simply because energy inputs per unit of 
output rises before the real price of copper rises, no conclusion to the effect that 
energy analysis has anticipated a relative price rise can he deduced. The an- 
alysis tells us nothing, for example, about technological change. It also assumes 
a decreasing quality gradient which, while it may he a sensible assumption for 
copper, is questionable for other resources. 

Finally, whereas the economic argument that current prices reflect future 
scarcity has some rationale to it, a rationale based on the maximising behaviour 
of resource owners, energy analysis offers a purely mechanistic interpretation 
of future scarcity based on the simple proposition that the exploitation of copper 
has led to the processing of leaner and leaner ores. To put it another way, what 
does energy analysis tell us that is not already contained in the fact, readily as- 
certainable, that copper concentration in ores has been declining over the years? 
If we reformulate the EA proposition as saying the declining ore quality is a 
sign of increasing scarcity we are stating the obvious as long as we ignore tech- 
nological change and the chance of higher quality ore discoveries. 

We can extend the analysis further. So far we have considered the case of an 
homogenous resource with no decreasing quality gradient, and the case where 
the quality gradient does decline. Now we consider the case in which we have sev- 
eral different types of energy inputs, say electricity from nuclear power, hydro- 
electricity and coal. In economic analysis different inputs are held to be effi- 
ciently allocated if the extra output (marginal product) produced by each last 
unit of input used to produce each output is the same. The marginal product of 
each input must be the same in all its uses. In the economist's language, the mar- 
ginal rate of transformation is common between all inputs. 

Now, different types of energy may be used in various ways — they can be 
used to supply energy directly, or they can be converted into other forms of en- 
ergy for indirect use (coal into electricity etc). As Turvey and Nobay have i>ointed 
out, 11 the marginal rate of transformation of factors in production (assuming the 
satisfaction of the conditions required to give an efficient allocation of resources) 
gives one economic measure of how one fuel type should be converted into each 
other. In cost terms the various fuel inputs should be valued at their marginal 
production costs. Where fuels are purchased by consumers the conversion factor 
is given by the rate at which the consumer is prepared at the margin to substi- 
tute one fuel for another so as to maximize his utility. This is the marginal rate 
of substitution and, assuming efficiency in the allocation of resources, is meas- 
ured by the marginal cost of the fuel to the consumer. 

What these concepts tell us is that in both production and consumption a therm 
is not necessarily a therm. Fuels have a number of attributes and heat con- 
tent is but one. Two fuels with the same heat content but other different at- 
tributes (in terms of cleanliness, transportability, etc) would have different 
marginal costs. In terms of economics the fact that two forms of fuel have the 
same heat content does not make those fuels identical. But in energy analysis 
the assumption of homogeneity (kilowatt hours are kilowatt hours regardless of 
how they are produced) obscures this important difference. The calculation of 
energy costs using the homogeneity assumption makes energy analysis irrelevant 
to the process of resource allocation now and over time. If. on the other hand, the 
homogeneity assumption is relaxed, energy analysis has no foundation. 

This point needs considerable emphasis since it lies at the heart of (he misuse 
of energy analysis, and is the foundation of its exaggerated importance. Only 



11 R. Turvey and A.R. Nobay. 'On measuring enerjry consumption'. Economic Journal, 
December lf)fi. r ). 



82 

by assuming homogeneity can EA proceed. But once homogeneity is assumed EA 
loses all relevance to resource allocation decisions. We may note that this pre- 
cludes EA from being used for virtually all of the purposes claimed by energy 
analysts. Examples are numerous. Thus it has been claimed u M that if the sly- 
erage thermal efficiency of power stations is 25% then 75% of the energy input is 
lost. But within the economic system consumers are demonstrating a preferance 
for a secondary fuel input over a primary fuel input. The price which they pay 
for electricity will reflect the opportunity costs of the inputs, including coal, 
used to make it. This is true irrespective of how much of the heat content of coal 
would be released in its alternative uses. 

In the market economy system if consumers demand coal as an input to make 
electricity they are saying that they value its use in this way more highly than 
in its alternative uses, even though in these alternative uses all the coal's heat 
content may be released. It follows that it is misleading to talk of the therms 
not directly converted into electricity as being 'lost'. Their alternative uses 
are assessed by consumers when the price system functions reasonably well. 

Another problem with the implicit homogeneity assumption can be illustrated 
using an example given by Chapman. 14 He sees one of the possible uses of energy 
analysis as being the ranking of alternative energy conservation investment proj- 
ects. Such projects are to be ranked simply in terms of the number of therms 
saved per £ invested. Now clearly this implies that all therms are equally 'worth' 
saving, whether they come from domestically produced coal, imported oil, or 
foreign enriched uranium. 

To appreciate some of the problems involved with this approach consider the 
following example. Suppose that expenditure of f 1 on the enforcement of speed 
limits led to a saving of 10 therms in reduced petroleum consumption ; that an 
expenditure of £1 on a law limiting the heating of public buildings led to a reduc- 
tion of coal consumption (used for electricity) of 11 therms and that an expend- 
iture of £1 on the development of a new gas cooker (of better efficiency) led to 
a saving of 12 therms. 

On the simple objective being used the last policy would be the best. But this 
choice would imply that not only would the alternative use values of the alter- 
native fuels be neglected, but, in addition, it would be assuming that saving a 
therm of imported energy was equivalent to saving a therm of domestically pro- 
duced energy. On the first of these points it is quite clear that if a government 
really wished to do this it could have a considerable amount of energy at little 
financial cost merely by introducing physical controls on the use of energy, 
eg rationing. But if such a policy were to consider only its costs of implementa- 
tion and the resulting energy savings in physical units, it would be ignoring the 
value of the benefits foregone due to the reduced consumption of energy. The 
problem then is simply that the saving of energy measured in physical units 
implicitly assumes a one-to-one correspondence of benefits foregone to energy 
(per therm) saved. In the market type economy there is no reason why this 
correspondence should exist. 

ENERGY ANALYSIS AS A NORMATIVE TECHNIQUE: ENERGY CONSERVATION 

So far we have tried to concentrate on what we might call the 'positive' claims 
of EA, claims which we feel have not been substantiated. We now turn to the 
wider claims for EA. These state that EA has some evaluative purpose. We are 
very much aware that energy analysts, in the main, have declared, in some 
cases repeatedly, that EA is not evaluative. Thus, Chapman declares 'Energy 
analysis does not tell anyone what they ought to do'. 13 Since energy analysis is 
merely an analytical technique this is what we would expect. Unfortunately, 
however, these same authors have then used this technique to make policy rec- 
ommendations. The same author has stated that energy analysis can be used to 
rank alternative energy conservation policies. 

Used by itself this is just what it cannot do. The ranking of alternative policies 
must be in terms of some specific objective function, and this function takes 
us away from the positive aspects of energy analysis into normative issues since 
this objective is necessarily not part of the analytical method. We might add that 
Chapman's denial of the evaluative role of EA is not supported by some of his 
colleagues. Hannon is quite explicit : 'In the long run we must adopt energy as 
a standard of value and perhaps even afford it legal rights'," (Our italics). 



12 P. Chapman and N.D. Mortimer. Energy Inputs and Outputs of Nuclear Power 
Stations, Open University Energy Research Group, Report ERG 005. 1974. 

13 P. Chapman. 'The Ins and outs of nuclear power'. New Scientist, 19 December 1974. 

14 B. Hannon, 'Energy conservation and the consumer', Science, vol 189, No 4197, July 11, 
J 975. 



83 

Fundamental to the choice between, or ranking of, alternative policies is the 
specification in an operational form of an objective function. In energy analysis 
this is also necessary in order that the boundary of the system should be de- 
lineated. In none of the papers on energy analysis that we have read is the ques- 
tion of the form and specification of the objective function discussed adequately. 
This may be the result of the (correct) view that energy analysis has no norma- 
tive significance. But, as has been mentioned, in many of these papers policy 
questions are discussed (and sometimes recommendations made) and so the 
relevant objective should have been stated clearly. In this connection it is im- 
portant to note that in economics the costs and benefits of particular actions 
cannot be defined or measured until the associated objective function has been 
specified operationally. Market price data are relevant in the pursuit of some 
objectives, while for others shadow prices must be used. Since the use of money 
values is sometimes recommended in energy analysis for the choice between 
alternative policies it follows that the associated objective must be specified. 

From the writings of a number of energy analysts it would appear that one 
of their prime concerns is with the question of energy conservation. They are 
particularly concerned to ensure that in the development of some energy source 
(e.g. shale oil or nuclear power) more energy is not invested that will be produced. 
It is therefore pertinent to enquire whether energy conservation can be con- 
sidered to be an (or the) objective. 

Expressed in this way the answer must be 'no' because it is non-operational. 
How much energy is to be 'conserved' over what time period? In what geograph- 
ical area? Since all methods of production involve the use of energy should 
the economic growth rate be chosen to maximise the rate of energy conserva- 
tion? Since even a zero growth rate involves positive production levels the re- 
quired growth rate would be negative. We presume this is not what is meant. 
Perhaps what is meant is that each productive process (when substitutes are 
available) should be selected so as to minimise the energy requirement. If this 
is the intention, then general, rather than partial, equilibrium analysis is re- 
quired and we note with interest that the development of satisfactory physical 
input-output tables would meet this objective. 

It remains the case, however, that objectives such as minimising the energy 
input of a given output are distinctively evaluative. It introduces the idea that 
energy as a constraint on economic activity is more important than any other 
constraint. If. for example, we selected policies on the basis of energy content, pre- 
ferring those with low energy input to those with higher energy input, we could 
easily find ourselves in a situation in which we would be adopting policies with 
high total resource cost. 

Certainly, energy conservation can be furthered by switching from energy-in- 
tensive products to non-energy-intensive products. The difficulty is to see why we 
need EA to further this end. Quite simply, the energy imputs into any economic 
process will have a price attached to them. This price will reflect the resource 
costs of supplying that input and included in these resources costs will be the 
energy inputs. In this way, the price of an energy input is built up from all the 
related previous processes. An energy conservation programme would require 
knowledge of how much energy costs will be saved by switching between 
products, information obtainable from a monetary input-output table just as 
readily as from a physical input-output table of the energy analysis kind. Further, 
the use of monetary measures would at least offer some indication of social 
preferences for the commodity switches whereas EA offers us no such guidance, 
as we have repeatedly pointed out. In short, energy conservation measures based 
on some index of energy input to commodity output implies an objective func- 
tion, but it is a function unrelated to consumer preferences and we see no 
justification for adopting EA as the appropriate technique when a preferable 
one exists. 

Chapman has stated explicitly "Thus if you want to adopt an effective energy 
conservation policy you can compare the costs of various policies (in £'s) with 
the amount of energy saved overall (in kWht's or therms or joules etc). This 
allows you to choose a best 'buy' in energy conservation." 10 Chapman claims 



15 P. Chapman. The relation of energy analysis to cost analysis, paper presented to 
Institution of Chemical Engineers working party on materials and energy resources. 1975. 



84 

that this choice cannot be made using economics because markets are imperfect 
and market-supplied data will be a poor guide to resource costs. In our view 
there are so many problems associated with this approach as to make it non- 
operational in the form outlined by Chapman. Further, although Chapman 
criticises economics he then (implicitly) uses it in his proposed method of choice. 

The first problem associated with this suggested method for ranking alterna- 
tive conservation projects is its total ambiguity with regard to the meaning of 
"costs". It is not clear whether cost refers to lifetime costs or to initial (invest- 
ment) costs, and whether these costs are to be aggregated in nominal terms or 
in time-discounted terms. In addition it is not clear whether these costs are to 
be given by market data or, given Chapman's strictures against economics when 
markets are imperfect, by the use of shadow prices. It is possible that what Chap- 
man has in mind is some kind of social cost-effectiveness analysis. But in that 
event the alternative policies should be compared in terms of achieving a speci- 
fied saving of energy, and the interpretation of costs as social opportunity costs, 
made explicit. 

All energy conservation measures will have a time dimension in the sense that 
they will take time to implement and their effects will endure. A problem which 
is immediately posed is that of determining the length of the planning period. 
This again involves the making of value judgements. Since the effects of any pro- 
posed policy will be uncertain a decision must also be taken on how to deal with 
risk and uncertainty. In particular it must be decided whether an error of, say, 
+50GJ in the estimate of energy requirements in any one year is to be con- 
sidered as being no worse than an error of — 50GJ with the exception of the sign 
difference. This would be equivalent to saying that the marginal utilities of 
equal size gains and losses were the same. Certainly we would doubt the value of 
energy studies which made no reference to the range of possible outcomes, and 
if possible with some probability estimates attached to them. 

To illustrate some of the problems involved with an energy analysis of alterna- 
tive conservation policies consider the following hypothetical example. For 
simplicity we assume perfect knowledge of the future and we will interpret the 
costs of the alternative policies to mean the initial costs. 



Option year 


1 
£ 


2 
£ 


3 
£ 


4 
GJ 


5 
GJ 


6 
GJ 


7 
GJ 


8 
GJ 


9 
GJ 


10 
GJ 


A 

B... 


50 

20 


100 
50 
80 


50 
130 
20 


100 
150 
50 


100 
150 
50 


100 
150 
50 


100 
100 
100 


100 
50 
150 


100 
50 
150 


100 

50 


C 


100 


150 



The question is then posed of how to choose between these alternative policies. 
Each policy involves the same expenditure (£200) and achieves the same saving 
in energy (700 GJ). However, the time distribution of both the expenditures and 
energy savings are different for the various policies. If we take Chapman's 
proposal at its face value we would have to assume that each of these policies 
was equally desirable. But this would not be the case if the policies were ranked 
using economic analysis. 

Firstly, from the economist's point of view neither the costs nor the con- 
sequences of these policies are the same for each policy option. This is because 
there is what is known as a time value of money, which simply says that equal 
nominal sums to be paid or received at different dates have different values 
when considered from the point of view of an individual or society. This means 
that before money sums occurring at different dates can be added together they 
must be re-expressed in terms of their values at some common date, such as 
the present or the terminal year of 'the policy. Whatever date is chosen a rate 
of interest must be selected. Now this involves many problems and the theoretical 
basis for this rate is the subject of dispute among economists. It would not be 
appropriate to go into these issues here, so we shall assume that the rate is 10% 
(equal to the test discount rate). Using this rate and discounting all costs to 
year 1, the costs of the three policies are £182, £173 and £189 for A, B and C 
respectively. 

The question must now be considered of whether a society would be indifferent 
between alternative energy savings policies which achieve the same total savings 
but with different allocations over time. In economics stress is laid upon the 



85 

time dimension in defining a commodity. This means that a unit of electricity in 
1976 is not considered to be identical ro a unit of electricity in 1980 or 1990. In 
economics it would not be valid to simply aggregate the energy savings occurring 
in different years. Before this aggregation can be made, as with the investment 
costs, all the energy savings must be expressed in terms of their equivalent values 
at some common year. Thus the energy units could be discounted to their equiv- 
alent year 1 values. Using 10% as the discount rate, the discounted energy savings 
are 401 GJ, 432 GJ and 366 GJ for policies A, B and C respectively. In economic 
analysis, for the data given policy B is preferred. But what do we know about 
this data ? 

Since in this example conservation polices are compared on the basis of joules 
saved per £ spent, an implicit assumption must be made that the price mechanism 
is working perfectly. If this assumption is nor made what significance attaches 
to the £ costs of the alternative conservation policies? Thus Chapman's strictures 
against the economist's market assumptions in their comparisons of alternative 
conservation policies apply equally to his own proposed method. However, 
economists do not always assume that markets operate perfectly and much 
modern work on project (and policy) appraisal involves 'the use of social cost- 
benefit and social cost effectiveness analysis where that assumption is not made. 

If by cost of the alternative policies is meant market determined initial costs, 
Then an important criticism of the suggested approach would be its total lack 
of attention to the costs incurred by a nation during years 4 to 10 inclusive. 
There is an implicit assumption that the recurring "cost" (however measured) 
is the same per joule of energy saved. But why should this be the case? 

Consider, as an example, the following two policy options hoth of which it is 
assumed give rise to the same total savings of energy and have the same initial 
cost. One policy involves saving energy by passing a law limiting the speed of 
road vehicles (with costs of new road signs and of the legislative process). The 
other involves expenditure on the thermal insulation of houses. The effects on 
producers and consumers per joule saved will be very different with these two 
policies. In the house insulation example consumers continue to enjoy the same 
or an improved level of home comfort and the initial costs of the policy are 
followed by a flow of energy savings which do not involve any reduction in 
consumer benefits. In the speed limit case, however, journeys will take longei 
increasing industrial costs etc and there will be generated benefits in the form 
of fewer accidents etc. It is clear that the effects per joule saved of different 
energy conservation policies could be very different. The only satisfactory way 
of proceeding would be to measure the costs of the alternative policies to include 
all the direct and indirect costs, and to define the costs in terms of some particular 
objective function. 

Earlier we discussed the homogeneity of energy assumption of energy analysis. 
When alternative energy conservation policies are considered this assumption 
is of crucial importance. This is because the way the energy conservation 
evaluative method is set up it is implicitly assumed that it is equally desirable to 
save 1 GJ of oil or 1 GJ of coal or 1 GJ of natural gas etc. But to look at the 
problem in this way ignores the differences in the relative reserve positions of 
the different fuels, the alternative uses which are available for those fuels (eg 
the use of oil as a fuel input or as an input into plastics), and of the geographical 
location of those different fuels. It seems to us that there would be a strong case 
for an identification within energy analysis of the effects on each different fuel 
of different conservation policies. 

ENERGY ANALYSIS AS A NORMATIVE TECHNIQUE I INVESTMENT APPRAISAL 

In some of their work energy analysts have adopted an investment criterion 
which economists know as the pay-back criterion. This criterion has played an 
important role in the energy analysis of nuclear power. Both Chapman and 
Mortimer 12 and Price 18 have calculated the number of years for individual 
nuclear stations and for programmes of such stations that will (on certain con- 
ventions and assumptions, and ignoring risk and uncertainty) elapse before the 
energv produced exceeds the energy consumed. The implication of their analysis 
is that the shorter is this period the better is the project. But this is not neces- 
sarily so, and there are a number of problems associated with the use of this 
criterion which must be clearly understood before it is used in the making of 
policy decisions. 



ia .T. Price. Dyanamic energy analysis and nuelear power, Friends of the Earth Ltd for 
Earth Resources Ltd. London. Dpcemhcr 1074. 



86 

An implicit assumption of this method is that the expenditure or saving of 
a nominal unit of energy has the same worth irrespective of when that expendi- 
ture or saving takes place. This criterion would rank the following two projects 
equally. The negative signs indicate a net energy consumption by the project, 
(on whatever measurement unit is chosen), while the positive signs indicate a 
net energy production. 



Project/year 


1 


2 


3 


4 


5 


6 


7 


8 


9 


10 


A 

B... 


—300 

—50 


—200 
—100 


—100 
—450 


600 
600 


600 
600 


600 
600 


600 
600 


600 
600 


600 
600 


600 
600 









The fundamental question which is raised is that of whether "society" is 
indifferent to exactly when energy is produced or consumed. This question im- 
mediately raises a number of complex issues, the most difficult of which is 
probably the determination of the relative weight to be given to a unit of energy 
production or consumption by different generations. The calculation of this 
weight is a matter of controversy among economists. But there is general agree- 
ment that its value declines through time and that it is less than one. This means 
that society prefers consumption (savings) which occur relatively early in time 
to those which occur relatively late. Applying this principle to projects A and B 
it would follow that society would prefer project B to A since, when allowance 
is made for time, it has lower costs but the same benefits. The first problem with 
the energy pay-back criterion is its neglect of the importance of time. 

Other problems involved with the use of this criterion include its failure to 
recognise explicitly the need for the normalisation of the lives and capital outlays 
of the alternatives projects. That is, how is a comparison to be made between 
two projects which have different estimated lives and investment outlays? 
Sufficient has been said to demonstrate the unsatisfactory nature of the energy 
pay-back criterion. 

CONCLUSIONS 

While most of the protagonists and practitioners of EA would probably agree 
that as an analytical method it is in its infancy and requires considerable refine- 
ment, the amount of publicity which has been given to the "conclusion" of some 
its papers makes it of paramount importance that the deficiencies and limitations 
of this method be widely understood. 

An example of the publicity given to one of its "policy conclusions" is the 
work of Chapman and Mortimer, and Price (cited above) on the optimal con- 
struction rate for a programme of nuclear power stations. At times, these authors 
have provided what must seem to the layman to be powerful arguments against 
the rapid build-up of nuclear generating capacity. Their concern has been to 
show that, ignoring all questions of the constraints in the construction industry 
on the maximum rate of construction, etc, that with some building programmes 
of thermal reactors, nuclear power "would always be a net consumer of energy : 
the more reactors we build, the more energy we should lose". 17 It is our contention 
that the policv consequences of the acceptance of this implicit policy recom- 
mendation are potentialv so serious that even at this stage of its development 
the methodology of EA needs to be subjected to detailed scrutiny and criticism. 

Our original' intension was to offer positive criticisms of and comments on 
EA in the hope that they would help with the further refinement of this method. 
At that time it was our belief that by providing a decision taker with more in- 
formation EA should be an aid to the taking of "good" decisions. Unfortunately 
as our study progressed the deficiencies in EA appeared to us to become more 
fundamental. Thus we must conclude that EA as now formulated and practised 
does not have any use beyond that which is currently served by some other 

ail As V wehave argued in this paper EA docs not ; (i) offer a method of evaluating 
projects, (ii) enable predictions to be made of changes in relative prices (either 
of the tvpe coal against oil. or energy against labour), or (iii) permi a choice 
to be made between alternative conservation measures. If energy analysis does 
not do any of these things, what does it do? We have been unable 'to And an 
answer to this question. If EA has uses not already adequately m rt by oth er 
techniques, it must be for energy analysts to demonstrate those uses by iboth a 



17 J. Price, op cit, p 24. 



87 

far more lucid exposition than they have so far provided and a direct comparison 
of EA and any other approach in a case study. It is our belief that the applica- 
tion of EA has run far ahead of the admirable motives that have produced it. 
In the absence of a convincing response to the challenge we have posed above we 
suggest that i't is a technique searching for a function. 



Net Energy Analysis — Is It Any Use? 
(By Gerald Leach) 

Net energy analysis began with two reasonable suspicions and an apparently 
simple method for testing them. The first suspicion was that as we turn to more 
diute and difficult energy sources the amount of "energy needed to get energy" 
will increase so that the net energy delivered will fall — perhaps in some cases 
to zero or less. In the long run this trend might be all-pervasive and set an 
ultimate physical limit on energy-based activities. The second suspicion was 
tha't traditional disciplines might miss these ominous trends because they either 
set narow system boundaries or use indirect units such as prices to meaure 
energy flow. Net energy analysis or NEA therefore proposed measuring "all" 
energy flows associated with energy supply (and conservation) technologies 
and measuring them directly in energy units of account. Since all energy fore- 
casting and policy issues are basically about such flows (though about other 
things as well) this procedure could sharpen all insights and decisions in the 
energy field. 

The idea had obvious appeal and caught on rapidly. From the start there 
were sceptics, chiefly economists, who often based their attacks on a mis- 
understanding of the humble aims of NEA as a descriptive science, believing 
they smelled heresy in the form of proscription and energy theories of value. 
But more recently skepticism and doubt have spread to net energy analysts them- 
selves, especially in recent months as the tide of studies produced a remarkable 
variety of methods, assumptions and "results" which could not easily be ex- 
plained away as merely the teething troubles of a new discipline. These worries 
culminated at the large NEA workshop held in August 1975 at Stanford, Cali- 
fornia 1 to compare and standardise precedures, where failures to resolve im- 
portant methodological issues were more common than consensus. 

In this article I take a critical look at NEA and suggest that the worried 
have good cause. At one level, I argue that as a practical tool for present day 
energy problems NEA is an elaborate sledgehammer for cracked nuts, adding 
little of importance to established energy studies. Nor does it have any special 
virtues as a longer term seer. At a deeper level. I suggest that NEA is plagued 
by methodological torments that cannot be resolved in any practically useful 
way, making it a Heath Robinson nut cracker. These are harsh conclusions and 
as a worker in the field I come to them reluctantly. The basic objections of NEA, 
like other comprehensive "look out" studies such as environmental impact analy- 
sis or technology assessment, are admirable. What I question here is how effec- 
tively NEA can ever support these fine aims in practice, while stressing that the 
question is a practical one in view of the widespread adoption of NEA by energy 
agencies, especially in the T T SA where Public Law 93-577 2 now requires a man- 
datory net energy analysis on new energy technology developments. 

TERMS DEFINED 

At. the outset I must emphasize that I am not discussing energy analysis in 
general, though some of the critique of NEA applies to this broader subject also. 
By estimating the total (fossil fuel) energy embodied in the final output of 
goods and services and thus capturing the often very substantial "hidden" indirect 
energy requirements for production, energy analysis has several most important 
uses. For example, it can map national energy flows in fine detail and thus help 
demand forecasts; it can identify energy intensive products: and it can say 
much about the inflationary impact of higher fuel prices. 



1 Draft Proceedings Report : Net Energv Analvsls Workshop, August 25-30, 1975 : In- 
stitute of Energy Studies, Stanford University. California. USA. 

2 Federal Non-nuclear Energy Research and Development Act of 1974. Section H(a) 
5. which reads : 'The potential for production of net energy by the proposed technology 
at the stage of commercial application shall he analyzed and considered in evaluating 
proposals.' 



88 

In contrast NEA studies only the energy requirements for energy products or 
savings. However, most of these have long been studied by traditional disciplines 
so that in a strict sense NEA adds only one new component: the previously 
"hidden" indirect requirements for materials, capital plant, non-energy products, 
services and the like. It is this "hidden" subsidy of delivered energy which is 
returned by the consumption sector to build up and operate the energy supply 
sector which NEA claims is important, which gives it its name, and on which, 
in my opinion, the value of NEA's contribution should be judged. 

IThis cardinal point and the terms I shall use are clarified by Figure 1 and 
the definitions below. The figure shows a generalised energy "module" which can 
represent any level of aggregation from a single stage of a fuel conversion chain 
(eg, a coal mine in Wyoming) to a whole national energy system and clearly 
shows the feedback loop / of energy subsidy which reduces the gross output of 
energy to a net amount available for the demand sector. The energy flows, which 
may be zero in some cases and for simplicity are assumed to be the sum of sepa- 
rate components, are : 




Resource 'lost' E t 




Final useful 
work and 
heat 



Waste E^ Waste heat E 



Figure 1. — Generalised energy system with main energy flows. 

Ei — principal energy input: eg, raw feed to a process stage or fuel extracted 
from a resource stock 8. 

E t — energy of resource 8 rendered unusuable by extraction : E,+Ei equals re- 
source reduction. 

E w — energy of principal input discarded as waste material : eg, coal spoil ura- 
nium mine tailings. E„ and many flows and flow-pairs E„ is strongly 
cost and technology dependent. 

Eh — waste heat discarded. 

E T — principal energy output returned to operate module or a previous stage in 
a chain : eg, gas pumped back into an oil well to increase fraction recovered. 

E — principal energy output crossing the system boundary. 

N — net output available, equals E —I. 

Et — final useful heat and work obtained from N: ie, after passing through end 
use appliances etc. 

/ — sum of external inputs crossing the systtm boundary, of which : 

I e — as direct fuels and electricity ; 

Im — as energy associated with non-energy inputs. These can be many and various 
depending on system boundary assumptions (see later), including energy 
for research and development, exploration, buildings, equipment, materials 
etc for capital and operational phases, government regulation, selling and 
advertising ; residuals management and decommissioning ; labour in various 
guises ; and the restoration of ecological side-effects. 

Note that I e is usually and I m is almost always, measured as a gross energy 
requirement or GER which records the total fossil fuel equivalent, a function of 
Ei. In many analyses this quantity is compared with system outputs E and A T . 
For brevity I will assume here that this direct comparison is legitimate, though 
it is not: a fact which raises some awkward data problems for NEA. 



89 

Now clearly this model is a great over-simplification of any real world system, 
even though it is able to record all energy flows. Above all it assumes away, or 
assumes that agreed solutions have been found to, four problems which in fact 
remain exceedingly intractable : 

1. how does one define the system boundary or, the same thing, know 
which external inputs / to count ; 

2. how does one allocate energies between joint products ; 

3. how does one 'add up' energies of different quality, or avoid hopeless 
overcomplexity by not doing so ; 

4. how when projecting an anlysis. e.g. over the 20-30 year lifespan of a 
facility, does one allow for the many cost and technology dependencies af- 
fecting flows and flow-pairs. This is particularly relevant for nuclear sys- 
tems where today's E* has a large but unknown potential as a resource S 
for tomorrow's technologies. 

These are the main issues I wish to discuss here. But before turning to them 
it is worth asking how relevant XEA is in the broad perspective of today's 
energy 'problematique'. 

ARE 'HIDDEN' SUBSIDIES IMPORTANT? 

Figure 2 shows the major (annual) fuel flows for a national system (excluding 
solar energy etc.). In fact by adopting three conventions 3 consistent with na- 
tional statistics the diagram is to scale for the UK energy system in 1968, though 
the quantities in dotted lines are only guesses. The feedback subsidy / is the 
latest estimate by Chapman.* 



Resources + reserves global 
and national 



total indentified \ 



Resource reduction 
I 
200' 



Energy extracted 
+ imports 



Gross output of demand 

Net output ol demand 



Supply technologies 



Final usclul energy 
(work, heat light etc) 
End use appliances 

I 35> 



Indirect energy purchases 
by energy supply industries 



Figure 2.— Main energy flows. UK 11)68. 

Two points about the diagram are immediately obvious. First. / is relatively 
trivial. The energy gain for the demand sector is N/1 = 24 while for resource 
use the inclusion of / to give a net rather than a gross efficiency (X/E instead of 
Eo/Ei) makes a difference of only 4 per cent. Second, there is a large reduction 
as one goes from resources on the left to useful energy on the right. The system 
is not only energetically inefficient but. needless to say in this journal, has a large 



3 The conventions are: (1) energy Is measured as heat content or enthalpy (1 kWli 
electrical=l kWh thermal) as in the 'heat supplied basis' tables of the Digest of UK En- 
ergy Statistics; (2) fuel and electricity transactions between energy industries are ac- 
counted for in the energy supply sector so that Ko is energy delivered to final con- 
sumers : and (3) nuclear fuels in Ei are counted as heat released and not as theoretical 
yield, making Etc for the nuclear sector zero with this one-year 'snapshot' view. Using 
OECD terminology Ei is 'total internal consumption' and Eo is 'total internal final con j 
sumption' less 'consumption by the energy sector'. 

4 P. Chapman, Fuel's Paradise: energy options for Britain (Penguin Books, London, 
1975). 



68-391 O - 76 



90 

potential for improvement over the next 10 to 30 years. Counting the whole 
panoply of conservation, energy income sources and technical measures there 
is undoubtedly a very large scope for providing the useful flows Et, N and E° with 
much smaller upstream counterparts towards the resources end. 

These potentialities and the uncertainties surrounding when they will be 
achieved and at what scale are, I suggest, so large as to render J an insignificant 
factor at present in any future-looking energy assessments. But the worry, of 
course, is that J might grow relatively in size and have some serious effect 
which, without NEA, would not be detected in advance. Is this worry legitimate? 
I think not. 

First, there seems no reason why NEA is particularly fitted to detect such 
trends. Sectoral forecasts of energy demand use available signals about the 
future to predict all components of E°, including J, for example, a rising demand 
for steel by the oil industry, and thus the energy associated with it, would be 
recorded as a higher energy consumption in the steel sector. It is not clear that 
NEA can provide better signals, nor that separating out I as a special compo- 
nent of demand has any particular virtues for forecasting. 

iSecondly, one has to ask what 'serious effects' an increase in / might have. 
It may or may not raise the financial or other costs of energy supply but these 
matters are outside the competence of NEA to answer since it deals only with 
energy flows. In terms of NEA the only effect is that to provide a given N there 
has to be a rise in E° and this may or may not lead to an accompanying increase 
in resource use (Ei or E»). So all that matters is what happens, if / increases, 
to the resource rations such as N/Ei and N/(Ei 4- E»). 

Let us now put some numbers into this argument. Table 1 gives some NEA 
data for a wide range of synthetic fuel sources. It shows that on the admittedly 
approximate estimates made to date N/ I ratios run from about 10 to 50, bracket- 
ting the present UK average of about 24. The most striking point about these 
data, though, is that in the 'worst' case — coal-to-oil — the inclusion of / makes a 
difference of only 10 per cent in the resource ratios shown in the last two columns. 
This figure is almost certainly inside the error margin for estimating the re- 
source ratios, especially when one takes future cost and technology changes into 
account. Also, for strategic energy questions the much more important factor 
is the overall change in energy outputs per unit of resource base compared to 
present day figures: how is the efficiency of resource consumption changing? 
Where these changes are large, as in most of the fuel conversions shown in the 
Table, 'rough cut' figures are normally quite adequate (given the inherent un- 
certainties in estimating them and in estimating / — see later) and can be ob- 
tained from almost any good text on modern energy technologies. 

TABLE 1— ENERGY INPUTS AND OUTPUTS FOR U.S. SYNTHETIC FUEL SOURCES 
Resources and inputs Outputs Ratios 



Ei+Es Ei Er I Eo N N-H N-rEi+Es Ei+E 



Shall oil: 

Sur face report: Room+pillar 

mine 195 

In situ report 282 

Coal-to-gas (high Btu gas): 

Western coal: Surface mined. 185 
Eastern coal: Deep mined 327 

Coal-to-gas (low Btu gas): 

Western coal: Surface mined.. 169 
Eastern coal : Deep mi ned 298 

Coal-to-oil: 

Western coal : Surface mined . 163 
Eastern coal: Deep mined 287 

Coal-to-methanol: 

Western coal: Surface mined.. 178 
Eastern coal: Deep mined 315 



121 
195 






6.5 
6.3 


100 
100 


93.5 
93.7 


14 
15 


0.48 
0.33 


0.51 
0.35 


180 
196 


0.72 
1.0 


2.0 
2.2 


100 
100 


98.0 
97.8 


49 
44 


0.53 
0.30 


0.54 
0.31 


164 
169 


0.91 
1.1 


5.1 
5.6 


100 
100 


94.9 
94.4 


19 
17 


0.56 
0.32 


0.59 
0.34 


158 
172 


0.70 
1.0 


8.4 
7.8 


100 
100 


91.6 
92.2 


11 
12 


0.56 
0.32 


0.61 
0.35 


173 
189 


0.73 
1.0 


2.2 
2.5 


100 
100 


97.8 
97.5 


44 
39 


0.55 
0.31 


0.56 
0.32 



Sources: Synthetic fuels commercialization program, vol. II: Cost/Benefit Analysis of Alternate Production Levels. Syn- 
fuels interagency task force to the President's Energy Resources Council, June 1975. Original data for standardized output 
Eo of 50,000 bbl oil equivalent per day or 110X10'* Btu per year adjusted to make Eo equal 100 arbitrary units. 

A second set of numbers also shows that resource consumption estimates are 
fairly insensitive to the inclusion of J inputs. Suppose that all new energy sources 
have an N/I ratio of only 5 : i.e., in this respect they are five times worse than 



91 

the present UK average and twice as bad as any of the sources in Table 1. but 
in all other respects are equivalent. Now consider three projections over 10 
years, all starting from the present UK pattern with net energy demand N 
growing at 3 percent a year. In the Base Case the present system merely expands. 
In Case A all additional supply comes from the new M/I=5 sources. In Case B 
the new sources also replace existing ones as such a rate that after 10 years they 
meet as much as 20 percent of total demand. This is, of course, a very extreme 
assumption. 

The effects on some important net energy parameters are shown in Table 2. 
Despite the severe assumptions and sharp falls in N/I for both Cases A and B. 
the resource consumption Ei rises by 'only' 5 and 9 percent respectively. This 
is not insignificant, but it is relatively small compared to present ignorance about 
more vital questions to do with energy flows. For example, electrification has 
persistently widened the gap in all Western nations between primary and de- 
livered energy Ei and Eo (when measured in enthalpies) and we have little idea 
how this is being reflected in final work and heat, the ultimate arbiter of energy 
demand and the proper basis for forecasting. 

If we know almost nothing about Ef in terms of resource use. especially 
by sectors, the refinements offered by net energy analysis, at least for the fore- 
seeable future and for national forecasting, seem rather luxurious. 

However, the severe assumption about future N/I ratios may be optimistic 
(though see Table 1) ; a decade is not long; and prudence, let alone Public Law 
93-577 in the USA, suggests that one ought to try to estimate the J inputs. I 
ask in the next four sections whether this is possible wih any acceptable degree 
of confidence for decision making or forecasting. 

THE BOUNDARY PROBLEM 1 

Net energy analysis is plagued by the problem of what external inputs / 
should legitimately be counted, which is the same as asking where one draws 
the boundary between energy supply and demand (see Figure 1). The solution 
adopted depends in part on the availability of data but mainly on which of 
two very basic and contrasting ideological assumptions one makes. It is this 
last point which has often led to the charge that NEA is a way of proving what- 
ever the energy analyst wants to prove. Yet the problem will not easily go 
away, which is awkward for a science aiming at reasonably accurate and com- 
parable quantitative results. 

The first approach to the boundary or counting problem is to draw the boundary 
between the energy supply system or facility being analysed and the rest of 
GXP as conventionally defined (the final bill of goods and services). This bound- 
ary is drawn automatically if all inputs / are counted using input-output methods 
since these are also consistent with the conventional definition of GNP. 

This approach may seem reasonable at first sight. However, as Bullard 5 has 
argued, it rests on the fundamental assumption that all activities within GNP 
are intrinsically 'good' and that so long as there are no external costs and future 
costs and benefits have been properly discounted, the flow of materials and energy 
through GXP should be maximised since it is not intrinsically 'bad' thereby to 
deplete resource stocks. More concretely, the energy used to build and run gaso- 
line stations, new towns for oil shale workers, energy using appliances, gas 
showrooms, and the Department of Energy or the Nuclear Regulatory Commis- 
sion and so on ad (almost) infinitum are seen as 'goods' within GNP and therefore 
not to be counted in J and charged as a cost on the energy supply sector. 

TABLE 2.-3 NET ENERGY SCENARIOS 



Ei 


Eo 


N 


1 


N/I Ei base case 


100 


69.8 
93.8 
98.8 
102.2 


67 
90 
90 
90 


2.8 
3.8 
8.8 
12.2 


24.0 


134 
141 

146 


24.0 1.00 

10.2 1.05 

7.4 1.09 



All cases: YearO... 

Base case: Year 10 

Case A: Year 10 

Case B: Year 10 

B C.W. Bullard, Net energy as a policy criterion (CAC document 154, Centre for Advanced 
Computation, University of Illinois. Urbana-Champaign. Illinois 61801, USA). 



92 

Naturally this view is strongly contested. At the concrete level, gasoline 
stations, new towns etc., would not be required but for the existence of the energy 
sector and are therefore not 'goods' but 'costs' to be included in the energy loop 
J. The fundamental assumption here is that welfare or utility is primarily a 
function of accumulated stocks and the flows needed to maintain them are costs 
to be minimised. 

At its most extreme, ,as represented by Howard Odum and the Florida school 
of analysts, 6 this paradigm forces one to capture all possible direct and indirect 
costs, including many remote multiplier and 'knock on' effects, such as the addi- 
tional energy associated with higher living standards for well-paid Alaskan oil 
workers, the energy to provide all social facilities and infrastructure for new 
energy developments, and all 'hidden' subsidies provided by natural ecosystem 
changes. With the latter one should, for example, count the loss of ecological 
capital due to soil run-off due to lessened vegetation cover due to poorer water 
quality and guantity supplied to farmers due to diversion of water to new energy 
facilities in the arid mid-Western USA. 7 As one would expect, even when only 
partialy followed through, this approach can generate large numbers for 7,° 
though with unknown but possibly large errors due to arbitrary truncation of 
the analysis. It also leads to a remarkable accounting method in which energy 
requirements for fuels increase in step with the prices of the fuels. 8 

Some may smile, yet the ecological and other effects are real and the methodol- 
ogy is a direct consequence of the paradigm assumption. The important question, 
though is how one compromises between the two paradigms. This is highly perti- 
nent since, at least on the evidence of the Stanford NEA workshop, 1 few analysts 
who adopt the GNP paradigm are willing to exclude some obvious components 
arising from the second. Chapman, 4 for example, deducts power used by electricity 
showrooms and offices from all electric facility outputs but makes no other 
obvious 'second paradigm' deductions ; and if one accounts for nuclear waste 
storage, why not also the energy associated with all residuals management, 
including health effects, of fossil-generated electricity? 

Until this problem is. settled by universal consent, NEA results will be arbi- 
trarily inconsistent, uncertain, and show large variations. (In the USA this will 
make Public Law 93-577 virtually unworkable since every published net energy 
statement will be wide open to reasonable objections and ensuing redrafts, delays 
and litigation). Yet consent itself demands an accepted ideological position on 
the two contrasting paradigms, which is hardly possible. While this meta-problem 
of course applies to all accounting procedures — including the use of monetary 
costs to capture all effects — it does suggest that NEA has no magic answer to 
some old dilemmas as it claims to have (see fourth sentence of this article, or 
Gilliland). 7 

THE BOUNDARY PROBLEM — 2 

Another serious boundary problem arises when one isolates any technology 
from others or from short-run cost changes. However, this isolation is necessary 
to NEA and its use in the decision process. This is because at whatever level the 
analysis, from major energy sources to variations in plant design, location, 
scale etc.. the analyst must work at a technology or plant level to gather engi- 
neering and material flow data. 

To illustrate this problem, consider Figure 3 which gives energy flows for an 
oil shale and refinery complex using data from Clark and Varisco, 9 who discuss 
the difficulties raised here. I shall ignore for the present the problems of joint 
production and energies of different qualities so evident in the diagram and accept 

6 For reviews of the Odum/Florida approach see H.T. Odum. Environment, Power, and 
Society (Wiley-Interscience, 1971) ; or more recently H.T. Odum, Ambio, 2, 220 (1973) 
See also for many references to specific studies the enthusiastic review of 'second 
paradigm' energy analysis by M.W. Gilliland, 'Energy analysis and public policy,' Science, 
189. 1 051— 1056. 

7 M.W. Gilliland. op. cit. 

8 M.W. Gilliland, op. cit. writes: "Imported oil at $2 per barrel has a net energy ratio 
of 30 to 1, while at $11 per barrel the ratio is 6 to 1" citing T. Ballantine. Net energy 
calculations of Northern Great Plains coal in power plants (unpublished paper. Uni- 
versity of Florida. Gainesville. 1974). This curious result arises because energy is seen 
as the sole driving force of the economy and creator of wealth : thus all extra dollars for 
fuels will incur pro rata energy expenditures somewhere ; and this allows money flows to 
be eauated directly with enprgy flows or consumption. 

e C.E. Clark and D.C. Varisco. Net energy and oil shale, paper to NSF (RANN) work- 
shop on net energy, University of California, La Jolla, January 1975. The data are adjusted 
to give 100 arbitrary units for output Eo. 



93 

the flows as given. Now the question is whether one can meaningfully write a 
simple net energy ratio such as N/I or even a net output quantity N for an isolated 
technology if the analysis is purely in energy flow terms. The answer appears to 
be no, as the following argument shows : 



I 11.4 



L electric 



9.0 
'GER' 



24 



Enthalpy 
3.0 



Ej 

143* 



GER 
0.67 



Mine, retort, refine, 
pipe to consumers 



Processed 
shale 



GER 
1.73 



Fuel 



Fuel oil 86.6 



Coke (unsold) 7.6 



Sulphur, 
ammonia etc 5.8 



y E o 10 ° 



Er 
26.7 



'Supply' I 'Demand' 



Figure 3. — Energy flows for a typical oil shale complex (35 gallons of oil per 

ton of shale). 

1. With the boundary shown there is a net ratio of 100/11.4=8.8 and a net 
output of 100-11.4=88.6. 

2. The price of electricity rises, making it worth diverting 9.0 units of output 
fuel to generate Ie (electric) within the complex: the net ratio leaps to (100- 
9.0)/2.4=38, though N remains at 88.6. In principle all / can be incorporated 
within the system boundary, giving an infinite ratio, while the ratio can be 
dropped well below the original 8.8 by supposing that returned fuel Er is sold 
as output and made up by Importing as I. 

3. For an isolated plant or technology ^V thus becomes the only meaningful 
output measure. Yet is has no meaning on its own since it is merely a scale- 
dependent quantity. X must be related to some more basic measure such as 
resource use, whether Ei or Ei-\-Es. Where a growth progamme is being analysed. 
A T must clearly be related also to assumptions about total demand for the energy 
product : indeed, it was the failure to do this that, among other faults, led the 
nuclear growth studies of Price 10 and Chapman and Mortimer 11 so badly astray 
(see Leach u and Brookes u among others) . 



10 J. Price, Dynamic energy analysis and nuclear power. Earth Resources Research Ltd, 
9 Poland Street, London Wl. England (December 1974). 

11 P.F. Chapman and N.D. Mortimer, Energy inputs and outputs for nuclear power 
stations (Report ERG 005. Energy Research Group. Open University. England. Decem- 
ber 1974). 

12 G. Leach. Nuclear energy balances in a world with ceilings (International Institute 
for Environment and Development. 27 Mortimer Street. London Wl, and 1525 New 
Hampshire Ave, Washington DC 20036, USA, December 1974 >. 

13 L.G. Brookes, 'Energy accounting and nuclear power.' Atom 227, September 1975 
(UK Atomic Energy Authority, London SW1). 



94 

While this may seem obvious, the consequences for NEA are serious. The 
resource parameters Ei, Es, Eo, Ew and Er are in most cases wide open to 
change due to short-run fluctuations of prices and interest rates, etc.. longer 
run changes in commercial or national resource depletion strategies, and tech- 
nological progress. While these change are difficult to predict they can have a 
large impact on the scale to which any technology is deployed and this scale will 
itself strongly influence the change in the parameters : no one bothers to improve 
tertiary oil recovery when this technique is not practised. Perhaps the most 
striking example of this circularity is in the nuclear field where the lifetime 
energetics of reactors built today depend strongly on the extent to which spent 
fuels (Ew) are used by future technologies through plutonium recycle and breed- 
ing, etc. More than this, though, the pace of present nuclear growth, and even 
of electricity growth as a whole, might be strongly affected if we could know now 
how rapidly these recycle technologies will develop or if they will be allowed 
to develop at all. 

The only way that NEA can handle these uncertainties is to give a range of 
energy performances for each facility or technology based on alternative assump- 
tions about the future. However, once one starts doing this it is a very short 
step indeed to having to model the total energy system, since many of the im- 
portant assumptions will depend on what is assumed to happen with other fuels 
and resources. Hence net energy analysis for most decision-making purposes is 
almost inseparable from longer term total-system energy modelling, with all its 
uncertainties. In which case, how useful is it for comparing specific technologies 
today? 

JOINT PRODUCTION 

The oil shale complex of Figure 3 demonstrates a relatively minor aspect of 
this thorny dilemma. How should Eo be counted? As enthalpy of the products? 
But why, since / is mostly as a gross energy requirement of GER? Then should 
the non-fuel products be given as the average GER for this production elsewhere 
in the economy? Or as the marginal GER for a drop in production elsewhere 
since the complex is now providing the materials? 

However, this dilemma is not minor in at least one cardinally important area 
for NEA, and nor are the questions so simple. The area is the mining and milling 
of uranium, which in tonnage terms is usually a joint product or a minor by- 
product (eg, Florida phosphates, gold in South Africa, copper at Jadugoda, 
India), while the contribution of these operations to the total / inputs for the 
nuclear system can vary between very large to insignificant depending on how 
one allocates energy costs between the mine products. 

To illustrate the important of this point, the original estimate of Chapman 
and Mortimer u for uranium from Florida phosphates was 12.557 x 10 6 kWht/ton 
UsOg. Applied to a SGHWR 1000 MWe reactor, this mining energy accounts for 
57 percent of the total energy invested to start-up, excluding content of the 
uranium itself: ie, 57 percent of the total J input for reactor construction and 
electrical equipment ; heavy water ; and uranium enrichment, conversion and 
fabrication for the first fuel load. Yet in their mining estimates, the authors ap- 
pear to have charged all or most of the energy to the uranium (0.013 per cent 
of the crude ore) and none or little to the phosphate (12 per cent ore grade), 14 
even though uranium extraction would be hopelessly uneconomic as the sole 
main product." 

Reviewing a conceptually similar problem — how to allocate energy inputs to 
output products in fuel industries — Chapman 18 has forcibly argued that alloca- 
tion by product weights or assigning all inputs to the principal product leads to 
logical absurdities. Allocation by price is more appealing but produces energy 
costs per physical unit that will vary in time and between different customers. 
He thus concludes that allocation according to the enthalpies of the products is 
the only sensible course. But none of these alternatives avoids logical absurdities 



14 The method of allocation is not made clear. However, the energy costs per ton of 
ore which Chapman and Mortimer charge wholly to uranium are similar to other esti 
mates for Florida phosphate production : see. for example, G. Leach. Energy and Food 
Production (International Institute for Environment and Development, London, August 
1975). 

16 To extract the uranium alone would have cost an estimated US $90-130 per kg U308 
in 1970 while in 1966 the byproduct cost of Florida uranium was US $22 per ker : M. 
Patterson, From bed to yellow cake — a comment on the energy cost of uranium (informal 
paper. International Institute for Environment and Development, London. July 1975). 

18 P.F. Chapman et al. The energy costs of fuels, Energy Policy, Vol. 2, pp. 2, 232-43. 



95 

in the case of joint production of fuels and non-fuel products where substitut- 
ability is zero, as with gold-uranium or phosphate-uranium — to give two of the 
simpler examples. Nor is there much point, as some might suggest, in allocating 
on the basis of marginal energy costs following a detailed process analysis : 
conditions would be so variable from mine to mine that aggregation to some aver- 
age figure would hardly be possible or realistic. 

In short, the joint production problem does not appear soluble by energy 
analysis or NEA in any universally acceptable way, while until it is major energy 
systems — eg, nuclear — cannot realistically be analysed. 

THE VALUATION PROBLEM 

Electricity is obviously more 'valuable' than coal, whether this value is meas- 
ured by price, social utility, or thermodynamic quality. Similarly peak load elec- 
tricity is not the 'same' as base load power, or why does anyone bother to build 
pumped storage schemes which are large heat or enthalpy sinks ? Energy analysis 
has long recognised this problem and the 'kippers and custard' dilemma of how 
one adds up energy flows of different quality, yet so far neither of the two con- 
trasting 'solutions' to it bodes well for turning NEA into a practical tool for 
decision making. (They may generate interesting theoretical insights, but this 
is not the issue I am addressing here) . 

One approach is to solve the add up problems by avoiding it altogether: the 
analysis should confine itself to displaying all flows and numbers separately 
(see, for example, reports of Working Groups I-A, I-B and I-C of the Stanford 
workshop 1 ). While the principle may be sensible, the practicalities, though, are 
daunting — especially with collecting the indirect input data by process analysis 
or input-output methods — for where does one stop the disaggregation into differ- 
ent energy types? And on what non-arbitrary criteria? Prices? The question is 
a serious one since any coarse sub-divisions — say into gas, three grades each of 
coal and oil, coke and other solids, plus base, mid and peak load electricity — will 
hopelessly complicate NEA as a vehicle for comparing technologies yet will still 
be too coarse for integrating the results with economic evaluations or reasonably 
sophisticated national energy models (see, for example, the 165 pathways and 
over one dozen 'energy types' of the French energy optimisation model 17 ). 

The second and contrasting approach is to insist on adding up so that results 
are useful, and doing so by employing a unique, physical unit of energy quality. 
Thermodynamic free energy or availability are the normal candidates. This 
approach too has its attractions as well as a strong theoretical underpining, but 
again leads to awesome operational difficulties. 

In this case the main difficulties arise from the large gap between the theoreti- 
cal concept and real world practices. While electricity, for example, may have 
a unique thermodynamic value as electricity, its practical value — ie, the thermo- 
dynamic potential actually extracted — depends entirely on its end use applica- 
tion which is, of course, both variable and largely unknown. Strictly speaking, 
NEA employing this approach must follow through to include all possible end 
use applications for energy and, perhaps, try to sum these : a most formidable 
task. Yet even when this is done, many important qualities of energy such as 
convenience, cleanliness, lower capital costs for all-electric installations and the 
like, escape the calculus. Monetary prices may have some arbitrary character- 
istics but do at least capture many of these differential qualities and, once set, 
are no longer open to question. 

CHECKS ON NEW SOURCES 

Despite these faults, one useful function of NEA has been to make rough checks 
on the effective limits to programmes or trends in developing 'new' energy sources 
or savings. Clearly, where energy must be invested before energy benefits are 
gained — whether in house insulation, nuclear power programmes, or solar elec- 
tric developments — over-rapid or massive growth can produce a long delay time 
before there are any net benefits. However, to date no analysis of which I am 
aware has raised any serious problems in this respect : all growths where net 
returns may be worryingly delayed or peak investment inputs worryingly largo 



17 D. Flnon, Optimisation model for the French enerpv sector, Energy Policy, Vol 2, No 
2. pp 136-51. 



96 

far exceed in pace and scale what is conceivable, due to the long lead times and 
inertia for energy developments or to constraints such as shortage of capital, 
labour, skills, materials, etc. 

For example, Chapman 4 shows that with typical UK conditions the energy 
investment for double glazing is paid back in 2.6 years by reduced fuel consump- 
tion with oil-fired heating systems (and slightly longer or shorter with other 
methods). A reasonably-paced programme to convert all existing 20 million houses 
by 2010, with a peak rate of 600,000 houses per year soon after 1990, produces a 
net energy deficit until about 1980 with a peak of only 1 per cent of present 
energy consumption for house heating. Thereafter the programme increasingly 
saves energy until at completion fuel used for house heating is three quarters 
of the present level. 

Similarly, crude data for solar-electric schemes pointing to pay back times of a 
few months 18 suggests that there can be no growth-critical problems with plausi- 
ble development programmes. The same applies to nuclear programmes when one 
allows for reductions in fossil-fired electricity and rejects the absurdity of in- 
definitely sustained rapid exponential growth. " 13 Shinnar " has also shown that 
tertiary oil recovery — a favorite target for net energy fears — must produce very 
substantial net energy gains on simple cost grounds for extraction costs up to 
around US $18 per barrel while a growth programme would have to be implausi- 
bly excessive to turn this into a deficit. 

While refinements of figures like these could be useful, they do not answer the 
crucial question of what one means by 'worrying' initial deficits in situations — 
as with all these examples — where one is capturing a new energy source, espe- 
cially if this is a perpetual income source such as solar or a 'perpetual' energy 
saving as with the insulation example. The proper evaluation of energy costs 
and benefits in these cases is extraordinarily difficult since it relates to major 
strategic decisions where some sacrifice in the short term provides long term 
gains, and hence to basically ethical questions of intertemporal equity and 
allocation. 

ULTIMATE LIMITS 

Finally, I shall discuss briefly the strong motivation underlying much net 
energy analysis (and energy analysis as a whole) : the quest for long term 
'points of futility' and ulimate sustainable limits to human activities. Slesser * 
has declared this to be the chief goal of NEA while Chapman's analyses of nu- 
clear systems running on low grade uranium ores were similarly motivated. 

The basic motivation has an obvious validity and appeal. At some point on 
the way down to 'zero grade' fuel sources it must take more thermodynamic 
potential to acquire the resource than it delivers. Identifying such point is 
crucial for longer term assessments. Similarly, energy analysis can 
use thermodynamic concepts to explore the minimum energies required to pro- 
duce things at various rates and thus measure actual performance against the 
theoretical best and perhaps set 'outer limits' to volumes and rates of production. 
The question, though, is whether one can ever forecast such points in the real 
world of certain, yet uncertainly certain, technical advances. 

An illustration from the nuclear must suffice to make the point. On the 
basis of 1950-70 data (when fuels were cheap) Chapman 4 has produced curves 
for the mining energy cost per ton of U 3 O g with ore grades ranging from 1 to 
0.001 per cent. The cost rises exponentially by about three orders of magnitude 
until at 20 ppm U 3 8 the mining energy exceeds the net yield that can be obtained 



18 Energy for mining silicon ore and reduction to Si is estimated as 30 kWht per kg with 
electric inputs counted at 30 per cent power station efficiency (R. Sambell and R. Davidge, 
Atom 215, 215-29). This gives 21 kWht/m 2 of 300 /on thick wafer. Electric inputs to a 
20-ribbon wafer pulling and purification machine are estimated at 19-20 kWhe/m 2 . or 
say 60 kWht (A. Mlavsky, Tyco Laboratories. Waltham. Mass., USA. personal com- 
munication). These are known to be the major inputs, which may perhaps be doubled to 
a total 175 kWht/m 2 to allow for equipment, materials etc. With a USA average year 
round insolation on a 45 degree south facing surface of 234 W/m 2 and 15 per cent con- 
version efficiency, annual electrical output from aim 2 unit is 308 kWhe (W. Morrow. 
Tech Rev December 1973. 31-42). Converting the inputs to 'electrical equivalent' gives 
a pay back time of 2 months. Various adjustments for more favourable solar regions, 
output losses for distribution, storage etc may alter this rough cut estimate by a factor 
of two or so. 

19 R. Shinnar, Net energy or energy analysis, paper to symposium on Economics of 
Scarce Resources. The City College of New York. 1975. 

20 M. Slesser. Discussion paper on an International Institute on World Energy Problems 
(20th Pugwash Symposium, Arc-et-Senans, France, July 1974). 



97 

from uranium in a present day (SGHWR) reactor. (E,„ credits and many inputs 
for the total cycle are ignored). Chapman concludes that 'Thus as far as our 
present reactor designs are concerned any source of uranium with a grade lower 
than 0.002 per cent is not useful in the sense that it does not produce a net 
energy output.' 

While this statement may he valid, of what use is it? And what is the use, 
even, of refining the estimates with more and better data? Present reactors may 
never need to run on such ores, whose grade is lower than the mine tailings 
from today's uranium production 15 : but we do not know this. If or when such 
more low grade ores must be used there will be nth generation technologies with 
better resource-using characteristics, or new technologies with a more accessible 
resource base : but we do not know the scale and rates of change. Meanwhile, 
there may be quantum leaps in the energy efficiency of uranium mining — eg, 
through high-rate leaching methods employing fast-acting bacteria that are 
even now being developed as a response to higher fuel costs in mining 1S : but 
again we do not know the trajectory of progress. 

These are not meant to be statements of despair, or of technological euphoria, 
but a plea for a certain humility. The future is opaque, a dark mirror, and no 
less to energy analysts than to the rest of mankind. Ultimate limits can wait on 
more urgent and closer concerns. 



APPENDIX III 

Analytical Methodology 

Leading practitioners of energy accounting recognized almost 
from the beginning (1973) that, with so many different possible ways 
to define systems and units, some sort of standardization in the field 
would soon become essential. Two international meetings were held 
under the auspices of the International Federation of Institutes 
of Advanced Study. Many agreements were reached, and areas of 
unresolved differences identified. The close connection of energy ac- 
counting with economics was the basis of the second meeting, to which 
a number of economists were invited. From these two reports, and the 
articles on methodology as it existed prior to the first meeting, one can 
see the juxtaposition of future potential value against current meth- 
odological difficulties. 



Use of Input-Output Analysis to Determine the Energy Cost of Goods 

and Services* 

(Robert A. Herendeen f) 

introduction 

Two-thirds of U.S. energy goes to uses other than direct personal consumption 
(that is, residential use or fuel for the automobile). In discussing an individual's 
total energy requirement, one must therefore pay attention to the indirect energy 
demand resulting from the demand for goods and services. In particular, one 
would like to know the energy impact of the production, maintenance, transporta- 
tion, and marketing of the whole spectrum of consumer products and to predict 
the energy requirements of hypothetical consumption patterns. 

This report discusses use of input-output analysis (I/O) to do this. The meth- 
odology is described in some detail, and specific reference is made to potential 
shortcomings of the approach. Complete results are given elsewhere; ' here I 
present several example uses, including calculation of the total efficiency of U.S. 
energy supply sectors, and the energy impact of agriculture and transportation. 

Input-output analysis is used for two reasons. First, there is a large body of 
data already available from the U.S. Department of Commerce, 2 at a fairly high 
level of disaggregation (367 sector economy). Second, the conceptual framework 
of I/O takes into account all steps in the complex manufacturing-sales chain. The 
latter point is important: results of this work indicate, for example, that auto- 
mobile manufacturing itself uses only about 6% of the energy needed to produce 
and market a car. 

The task of converting the existing I/O data, which are in terms of dollars, to 
energy is nontrivial because energy is sold at different prices to different customers. 
For electricity, for example, there is a factor of 6 in variation. 

input-output theory: deriving energy "cost" 

dollar relationships 

The data for I/O are dollar sales per year between sectors of the economy. (Of the 
367 sectors, 5 are energy producers: coal mining, oil and gas wells, petroleum 
refining, electrical utilities, and gas utilities.) One then assumes linearity and time 
invariance to write 

n 

x,=2^,x,+y„ (i2i) 

where 

X t = the total output (dollars) of sector t, 
Fj=the output (dollars) of i sold to final demand, 
A a = constants, obtained empirically from the data: 
Aa = [sales (i — » j) /total sales j] for the study year. 

The assumption of constant An is that of technological constancy (measured in 
dollars). It is assumed that Equation 12.1 holds for arbitrary final demands. In 
matrix form, 

X=AX+ForX=(l-A)-iy. 

The matrix (I — A) -1 for the United States economy has been published (Bureau 
of Economic Analysis, BEA, 1969) for 1963. The 1967 results are expected this fall. 



•This work was begun while the author was a staff member of the Oak Ridge Environmental Program, 
Oak Ridge National Laboratory, Oak Ridge, Term., 37830, and continued at the Center for Advanced 
Computation, University of Illinois, Urbana, 111., 61801. At both institutions it was supported by grant 
from the National Science Foundation RANN program. 

tCenter for Advanced Computation, University of Illinois, Urbana, Illinois. 

1 See Herendeen (in press). The results, with less amplifying material, are contained in Herendeen (1973). 

2 See Bureau of Economic Analysis (1969). The BEA has published several amplifying articles, Bureau of 
Economic Analysis (1969 November, 1971 January, 1971 August, and 1972). The complete tables are also 
available on tape from BEA. 

(101) 



102 

12.2.2. INTRODUCING ENERGY 3 

To convert to energy terms, let 

n 

Ei^Zj/Eik + Eiy, 
where 

Ei = total energy output (Btu) of energy sector i, 
E ik = energy sales (Btu) from i to k, 
E iu — energy (Btu) of type i sold to final demand. 

Since 

B -=(S*Hl7§ [(| - A) - ,u ' y '' 

k = \ 1 = 1 A k \Y it 

TABLE 12.1— SALES FOR A YEAR OF A 3-SECT0R ECONOMY « 

Crude Refined Final Total 

oil petroleum Cars demand output 

Crudeoil 10(40) 10 

Refined petroleum 5(5) 5(5) 5(5) 25(25) 40(40) 

Cars 20 20 

1 Figures in parentheses are energy sales; the other are dollar sales. 

(a) R ik = E ik /X k 

( u\ a _\ E JYi, i=k energy sector, 
W a*-*-| 0) otherwise. 

Then 

E=[R(\ - A)~ 1 + S]y = tY, (12.2) 

where e is the total energy matrix. It is a 5 by 362 matrix * and €,, gives the 
total output (Btu) of energy sector i required for the economy to deliver a dollar's 
worth of project j to final demand, $1 FD (j). It should be noted that in con- 
structing R, one is constrained to choose the energy flows appropriate to the 
period and sector definition use in constructing A. In a real economic system R 
and A are not independent. 

As an example, consider the 3-sector economy, whose sales for a year are given 
in Table 12.1. 

T 1/4 01 T7/6 1/3 1/12-1 

A= 1/2 1/8 1/4 ; (I - A)"' = 2/3 4/3 1/3 , 

LoooJ Loo 1J 

R= r° x °i = r 2/3 4/3 l/3 ~i+r o ° °i 

Ll/2 1/8 1/4 J' |_2/3 1/3 l/3_r|_0 1 OJ 

Additional information can be obtained from this approach. For example, how 
much energy is used to make the (say) steel contained in a car? This can be ob- 
tained from individual terms in the matrix product R(l — A) -1 . #,•*[(! — A) -1 ]*, is 
the energy type i supplied to k in order for k to supply enough of its output for 
the economy to deliver $1 FD (j). 

For the example, writing Equation 12.2 in terms of uncollapsed sums yields 

f 2/3 4/3 1/3 "I ro 01 

C L (7/12+ 1/12) (1/6+1/6) (1/24+ 1/24+ 1/4) J + Lo 1 oJ' 

3 Reardon (1972) has done similar work for a 35-sector breakdown. 
* Five of the sectors use no energy and are deleted. 



103 

Thus, of the total of 1/3 Btu of refined petroleum required to deliver $1 FD 
(car), 1/24 Btu is used by crude, 1/24 Btu by the refined petroleum sector itself, 
and 1/4 Btu by cars. Table 12.2 shows an application of this for the automobile. 



12.2.3. A PROBLEM WITH PRIMARY ENERGY SECTORS 

The approach outlined so far does not assure that the amount of refined pe- 
troleum required per $1 FD (j) wm * not exceed the amount of crude. In fact, it 
is possible to construct an economy that does this, and yet is monetarily and 
energetically realistic. The problem is one of allocation — the present method does 
not use the refined petroleum pricing data (there is a price implied in R i; ) to allocate 
crude; it merely uses the dollar flows. Explicit treatment is needed. 

The method I suggest transfers the crude sold to refiners to the refineries (or in 
general, from primary to the consuming secondary industry), "letting" the 
refined sector distribute it, and then allocates this use back to crude. In the ex- 
ample, this requires letting ftorude—ref^O to produce a new R(r) (reduced) and a 
new e(r) : 
e(r) = [R(r)(l - A)-» + S]. 



The new total energy coefficient (called T) has a new crude row: 

T'crude- > j == Corude-»j('") + (■^crude-»ref/-^ref)€ref- > j(r) . 



(12.3) 



To demonstrate, change the refined petroleum Btu sales in the example to 5,5, 
10, aiSd 20 Btu. (Note that this implies new prices: a new economy.) Without the 
correction, 

= T2/3 4/3 1/3 "I TO 01 
€ l_2/3 1/3 7/12j + |_0 4/5 0J 



and shows that cars require more refined than crude/ 
But 

W ~L2/3 1/3 7/12_J + Lo 4/5 0j' 
TABLE 12.2— ENERGY BREAKDOWN FOR MOTOR VEHICLES (59.03), 1963 









PCINVB2 












Crude oil 








T0TPCP3 






and gas 


Refined 


Elec- 




total 


Row and description » 


Coal 


extraction 


petroleum 


tricity 


Gas 


primary 


1. Food . 


0.18 


0.47 


0.81 


0.28 


0.24 


0.36 


2. Construction 


.03 


.80 


1.82 


.11 


.08 


.53 


3. Textiles 


.61 


.56 


.57 


1.32 


.58 


.50 


4. Paper and lumber 


1.61 


1.77 


1.64 


1.31 


1.95 


1.60 


5. Furniture 


.01 


.01 


.01 


.02 


.01 


.01 


6. Chemicals 


6.25 


12.24 


10.34 


10.44 


12.68 


9.53 


7. Rubber 


1.35 


1.00 


.96 


1.91 


1.07 


1.00 


8. Leather 


.01 


.01 


.02 


.01 


.01 


.01 


9. Stone, clay, glass 


.86 


3.12 


.72 


1.61 


5.03 


2.24 


10. Primary metals 


65.11 


34.29 


24.61 


34.78 


43.02 


41.24 


11. Fabricated metal 


.62 
.87 


1.28 
1.39 


1.32 
1.59 


1.70 
2.32 


1.31 
1.31 


.96 


12. Machinery... 


1.09 


13. Instruments 


.69 


.84 


.65 


1.86 


1.02 


.69 


14. Transportation equipment 


15.12 


11.01 


8.14 


25.23 


13.64 


10.92 


15. Transportation services 


.48 


11.33 


24.78 


.69 


1.92 


7.52 


16. Mining (metal, stone, fertilizer). . . 


1.45 


1.70 


1.46 


2.34 


1.96 


1.47 


17. Coal mining 


.89 


.24 


.35 


.80 


.16 


.40 


18. Crude gas 


.07 


2.02 


.16 


.27 


.42 


1.33 


19. Petroleum refining, and so on 


.18 


5.39 


9.64 


.50 


2.50 


3.55 


20. Electric utilities 


1.52 
.04 
.05 


.76 

2.31 

.26 


.34 
.24 
.43 


6.01 
.03 
.07 


1.11 

3.94 

.14 


.74 


21. Gas utilities 


1.52 


22. Water 


.18 


23. Wholesale and retail trade 


.38 


2.49 


4.43 


1.52 


1.18 


1.69 


24. Finance insurance and business... 


.36 


1.17 


1.16 


1.41 


1.23 


.82 


25. Medical 




















26. Education 


.04 


.09 


.10 


.16 


.08 


.06 


27. Advertising 


.01 


.07 


.08 


.02 


.06 


.04 


28. Radio, TV, communications 


.11 


.32 


.44 


.39 


.24 


.23 



1 These sectors represent an aggregation of 367 into 28. Some sectors have been omitted, hence, column sums are less 
than unity. 

> (PCINVB),i is the percentage of the total energy of type i needed to make a motor vehicle that was used in sector j. 
Thus, 65.11 percent of the coal needed to make a motor vehicle was used by the primary metals sectors. 

• (TOTPCP)j is the corresponding percentage of the total primary energy. Of the primary energy needed to make a 
motor vehicle, 41.24 percent was used by primary metals. 



104 

and, since E crude ^ Te f/E ref = l, 

T2/3 1/3 7/12-1 ro 4/5 01 
L2/3 1/3 7/12.T |_0 4/5 Oj' 

Energy is now "conserved," while some crude is allocated to final demand. 
In the Department of Commerce I/O matrix, there are five energy sectors: 

Index : I/O Sector Title 

1 7.00 Coalmining. 

2 8.00 Oil and gas wells. 

3 31.01 Refined petroleum. 

4 68.01 Electric utilities. 

5 68.02 Gas utilities. 

The e(r) is obtained with the following energy flows set equal to 0: 

1 — > 4, coal — * electrical utilities, 

2 — > 3, crude oil and gas — > refined petroleum, 

2 — > 5, crude oil and gas — * gas utilities, 

3 — ► 4, refined petroleum — > electric utilities, 
5 — > 4, gas utilities — ► electric utilities. 

By the same reasoning as earlier 5 

r«=SCa«H(0, orT = C[R(r) (l-A)~i + S], (12.4) 

k 

where 

C,, = l, i=l, 2 . . . 5, 

Cu = EiJE if 

C 23 = E 23 /E 3 , 

C 2i =(E 23 /E 3 ) (E 3i /E i )-\-(E 25 /E 5 ) (E b <E A ), 

C 25 = E 25 /E 5 , 

Cu= E34/E4, 

C 5i = Es i /E i . 

2.3. LIMITATIONS OF THE i/o APPROACH 

Some of the limitations result from data-handling problems. Several, however, 
are conceptual and derive mostly from economic conventions used in constructing 
the published I/O tables: 

1. Input-output data are subject to inaccuracies from (1) lack of complete 
coverage of an industry, (2) restriction of information for proprietary reasons, (3) 
use of different time periods for different kinds of data. Errors in A may generate 
disproportionate errors in (1 — A) -1 . We are actively pursuing this last point. 
Preliminary analysis shows the published (1-A) -1 to have a condition number 
of 5; this is a rough index of the multiplication of error. 

2. The use of dollars rather than physical units to express physical dependencies 
is risky. For example, aggregation can combine two processes whose energy 
intensities differ widely in the same sector. And dollar economies of scale may be 
implicit in the A a, whereas there would be no (or little) corresponding effect in 
physical terms. We are also concerned with this question. 

3. There is the problem of transfers. BEA's I/O sector definition is based on the 
establishment, rather than activity. For example, if those establishments that 
produce primary aluminum also produce aluminum castings (amounting to less 
than 50% of total sales), the primary aluminum sector is credited with their 
complete output. The secondary output is transferred to the aluminum castings 
sector, that is, treated as a sale. The corresponding inputs are not transferred. 
This means that the dollar output corresponding to production of these aluminum 
castings has been counted twice, but the energy only once. The fraction transferred 
varies from sector to sector, so that R„ require a correction, details are in Heren- 
deen (in press). 



' This is still an approximation, as I have not accounted for electricity used by refineries, and so forth. 
These energy flows are small compared with those included in Equation 12.4. 



105 

4. A problem arises in capital goods; these are not considered part of the inter- 
industry transactions but are listed as sales to final demand. Conceptually, I 
could consider the energy to make a steel forming press owned by an auto manu- 
facturer to be as valid an energy contribution to my car as that used to make the 
steel, but this is not the convention used in I/O. One must learn in detail how this 
is handled, since capital is defined differently from industry to industry and sways 
with the winds of the tax laws. Given reliable figures, one can separate the capital 
for depreciation from that for expansion, and allocate it (as its energy input) to 
the actual consuming sector. We are actively studying this question. 

5. Final demand is measured in producer's, not purchaser's, prices. Since the 
I/O sectors include wholesale and retail trade, it is possible to make the conversion, 
including the energy requirements implied in the markup (as has been done for the 
automobile in Section 12.3). But for direct use by consumers, it is desirable to 
convert beforehand to producer's prices. Data on markups are available from the 
Department of Commerce, and we are now using them. 

6. Input-output coefficients change with time, yet we hope to use the results to 
predict the consequences of hypothetical consumption patterns. Can one quantify 
their loss or reliability with time? This is a major point, for which much work is 
needed. We shall be converting the 1967 I/O table to energy terms for comparison 
with 1963, and we are starting sensitivity analyses to determine which coefficients 
are most critical. Our feeling is that energy use coefficients, on the whole, change 
faster than others. 

7. Input-output sectors are often too broad. For example, do buses require less 
energy than cars for manufacture (per dollar)? To answer this, we are currently 
working on two tacks. One is to gather additional data to allow disaggregation of 
the "motor vehicles and parts" sector into two. This requires construction of both 
a row and a column, and an inversion of the new n+1 degree matrix. A second is 
to use the identity 

(|-A)-i = l + (l-A)-iA 

= I + A + (I-A)- 1 A2 
and so on (12.5) 

as the basis for an approximation. For example, for the capital energy cost of a 
mass transit system, we could adapt Equation 12.5 thus: 

AE=AE' + CR(\-A)- i AY, 

where 

AE=the total energy cost, 

AE' = direct energy used, 

and Ay=the bill of the input goods, 

stated compatibly with I/O categories. Intuitively, this is just a first-order vertical 
analysis (that is, an intuitive step by step tracing of inputs) except that the energy 
content associated with the inputs is given by the total energy coefficient. This is 
called a "hybrid" approach. 

12.4. WHAT IS THE ENERGY NEEDED TO MAKE A PRODUCT OR PROVIDE A SERVICE? 

The total energy coefficient 7\, allows an answer when the question is modified 
to read "to supply a product to final demand." One can allocate all of the U.S. 
energy budget this way (that is, E=TY). But this scheme gives absurdly low 
weight to sectors that sell little of their output to final demand (for example, pri- 
mary metals). Trying to assign an energy to this kind of sector is difficult. An addi- 
tional ton of aluminum can be required as a result of many different changes in 
final demand (for example, for more milk!), and each will have a different energy 
requirement; there is no unambiguous energy impact of the additional aluminum 
production. The argument below helps to resolve this. 8 



8 1 am Indebted to Clark Bullard of the Center for Advanced Computation, University of Illinois, for this 
approach. 



68-391 O - 76 - 8 



106 



Let a sector receive (1) direct energy from energy sectors or the "earth" (if an 
energy sector), (2) embodied energy by virtue of purchases of processed inputs, and 
let it distribute (1) actual energy (if an energy sector) and (2) embodied energy in 
its sales. 




Here we will assume only one energy sector, but generalization is easy. For the 
energy sector m, J^earth-m^^m-out; these quantities are zero for the other sectors; 
E t , = unless i=m. The model assumes that embodied energy content per dollar 
e, is the same for all sales by j and does not differentiate sales to final demand. 
Energy balance requires 

n 

ejXj — 2_j C,X ij = Eij + -dearth _ -Sj-out- 

fel (12.6) 

The right-hand side is the energy dissipated in j; call it r/, 






(12.7) 
e = J R(|-A)-i (12.8) 

The embodied energy coefficient is thus identical to the total energy coefficient 
obtained before, but now it is associated with all sales of j, not just those to final 
demand. If we used total outputs to allocate energy, we would obtain a total 
energy several times greater than the U.S. energy use, since there is much double 
counting. One obtains this picture of the economy: 

energy enters through the primary sectors, a greater quantity of embodied energy 
circulates, and an amount equal to the input leaves via final demand. 

It is possible to allocate on the basis of total output, but then a disclaimer is 
needed: "The energy necessary to supply the production of widgets last year is 
given by the total output times the total energy coefficient. Many of these widgets 
would not have been produced were there not also production of gizmos which 
require widgets as inputs. The energy needed to produce gizmos thus includes 
that to produce the required widgets, and allocating energy this way double 
counts." 

12.5. SELECTED APPLICATIONS 

The total energy coefficients for the 1963 I/O table are listed in Herendeen (in 
press) . (The main task of that work was to obtain the energies used by the various 
sectors.) Here I present several uses for the results. All these are subject to as yet 
unquantified errors, and the capital problem referred to. On the average, my 
feeling is that the numbers for 1963 are good to ±15%. 

12.5.1. BREAKDOWN OF THE ENERGY NEEDED TO MAKE A MOTOR VEHICLE, 1963 

(TABLE 12.2) 

This uses the method of Section 12.2.2, which illustrates how the energy to make a 
product is built up from the energy used by all the suppliers of components of 
that product. For the auto, over 40% of the total is used by the primary metals 
sector, and only about 1 1 % is used by the transportation equipment sectors. 



107 

12.5.2. ENERGY EFFICIENCY OF THE ENERGY SUPPLY SECTORS 

The total primary energy coefficient T, (prim) is denned as the sum of the coeffi- 
cients for coal, crude oil, and gas extraction and a portion of electricity produced by 
hydro, converted at the going heat rate. Thus, for energy type i, the fraction 
(total primary energy) /(l Btu of type i to final demand) is 



(!2 , T ki yS ii =T i (pnm)/S, 



(12.9) 



where £' is the sum referred to. Table 12.3 lists the results for 1963. 
k 



TABLE 12.3.— INVERSE ENERGY EFFICIENCY (IEE) OF THE ENERGY-PRODUCING SECTORS 
[The total primary energy required to deliver 1 Btu of energy of various types to final demand, 1963. See sec. 12.5.2) 



Title 


I/O sector 


IEE 
(Btu/Btu) 


IEE 1 

corrected 
for imports 


EE2 
(reciprocal) 


Coal mining 

Petroleum refining 


7.00 

31.01 


1.024 
1.082 
3.870 
1.134 


1.024 
1.208 
3.870 
1.169 


0.977 
.828 


Electric utilities ... 


68.01 


.258 


Gas utilities 


68.02 


.855 









1 10.4 percent of refined petroleum and 3 percent of gas were imported in 1963; the energy to produce them was ex- 
pended outside the U.S. economy. 

J Since these results are based on producers' prices, additional energy would be expended in delivering coal and refined 
petroleum to most consumers. For electric and gas utilities, this is not a problem. 

12.5.3 AGRICULTURE AND FOOD 

1. Agriculture. Ten sectors (1.01-1.03, 2.01-2.07) of the 367 are identified as 
agricultural (Bureau of Economic Analysis, 1969), and the energy to support their 
production can be obtained by summing the total primary energy needed to pro- 
duce each sector's output minus that sold to other agricultural sectors; that is, 



Vj Ti (prim) fl-SlA 



(12.10) 



where G is the set of agricultural sectors. This is the energy needed to allow the 
aggregated agricultural sector to ship to everyone else (mostly the food processing 
industry). For 1963, this was 2.2 X 10 15 Btu, which is ~4.4% of the U.S. energy 
budget for that year. The direct energv use for these sectors (with electricity con- 
verted to primary by a factor of 3.9 '(Table 12.3) was 1.2X10 15 Btu (this direct 
figure agrees well with recent ones from Perelman, 1972). Agriculture thus required 
about as much energy to produce its nonenergy inputs such as fertilizer, equip- 
ment, and so on, as it did for direct use. 

2. Food and kindred products. The same analysis is applied to the food sectors 
(I/O 14.01 to 14.32) to yield 3.6X10 15 Btu, which is ~7.2% of the U.S. total. 
Much of the agricultural energy has been counted again in this figure. The total 
figure can be compared with the calorie content of the food, approximately 
0.8X10 15 Btu (Perelman, 1972). The embodied energy is 4H-times the food 
energy. 

3. The energy cost of protein. Several I/O sectors are for high-protein foods, 
and the energy cost to produce the final processed, packaged product has been 
calculated as a function of the protein contained. Table 12.4 lists results for four 
sectors: meat products, cheese, milk, and fish. It seems that in 1963 fish was a low- 
energy protein source and milk a high one. For comparison with the work of others, 
recall that this is the total direct and indirect energy. 



108 



12.5.4 TRANSPORTATION 

The direct fuel use for all transportation today is about a quarter of all energy, 
but much more is required to supply the support: transportation equipment, 
roads, and so on. 

1. Energy impact of the automobile, 1963. 

Table 12.5 is an application of the method described in Section 12.4. It involves 
multiplying the appropriate total primary energy coefficients times the corres- 
ponding final demand expenditures associated with the private automobile. In 
1963, the average new domestic automobile was worth $1,890 in producers' prices, 
which implies a total energy cost of 132 X 10 6 Btu/car (the equivalent of 5 tons 
of coal). This figure agrees very well with that obtained by Berry and Fels (1972). 

The final demand expenditures associated with the automobile were 12.4% of 
the GNP, but accounted for 20.7% of the nation's energy budget. Only 57% of 
the automobile energy was for direct use as fuel. These statements can be made 
without disclaimer since all pertinent expenditures are part of final demand. 



Title 



T(primary) 1 
I/O sector (kcal/$) 



Producers 
price 2 
($/lb) 



Percent 
proteins 



Protein 
energy * 
(kcal/lb) 



Energy/ 
calorie » 
content 



Meat products 

Cheese, natural and processed. 

Fluid milk 

Fresh or frozen packaged fish.. 



14.01 


14, 200 


0.50 


22.0 


32, 600 


14.03 


15,600 


.30 


25.0 


18,800 


14.06 


15, 000 


.12 


3.5 


51, 200 


14.12 


10, 100 


.35 


20.0 


17, 700 



6.3 
2.6 
6.1 
6.5 



i 1 kcal = 4.0 Btu. 

2 Bureau of the Census (1971). Figures are approximate. 

a (Bowes and Church, 19 7 0). 

4 Marketing energy not included; on the average it would increase figures by 10 to 15 percent. 

s Total calories including nonprotein materials. 

TABLE 12.5— ENERGY IMPACT OF THE AUTOMOBILE, 1963 » 



Dollar flow 
(10°) 



I/O sector 



I/O coeffi- 
cient b 
(Btu/?) 



Energy, 10 12 
Btu 



Percent of 
total 



Gasoline: 

Production 

Refining 

Retail markup • 

Oil: 

Production 

Retail markup 

Auto: 

Manufacture 

Retail markup 

Repairs, maintenance, parts.. 

Parking, garaging 

Tires: 

Manufacture 

Retail markup 

Insurance 

Taxes (highway construction) 

Total 



» The analysis is described by Hirst and Herendeen (1973). The numbers here differ somewhat since the calculation 
there was for 1970. 

* From Herendeen (in press). These are expressed in total primary energy, including a thermal equivalent of hydropower. 
See sec. 12.5.2. 

«• Figures obtained from American Petroleum Institute (1971 ed.), pp. 306, 307, 322 and Bureau of the Censes (1971, 
pp. 536-537). There were 69x10 6 autos registered in 1963, and 7.64X10 6 produced domestically at an average producer's 
price of $1,890; 0.41 Xl0« were imported. 

" See table 12.3 (Btu/Btu). 

• Retail gasoline markup and taxes from American Petroleum Institute (1971 ed., pp. 458-459). 
/ Oil/gasoline ratio equals 128 on a Btu basis, from Bureau of the Census (1971, p. 537). 

i Markup of oil and tires assumed to be 40 percent of purchaser's price. 
•12.4 percent of GNP. 
t 20.5 percent of total. 



« $5. 86 


31.01 ... 

31.01 

69.02 

31.01 ... 
69.02 

59.03 
69.02 
75.00 
75.00 

32.01 
69.02 
70.04 
11.04 


"<*"6."268" 
32, 700 

"32,766" 

70, 000 

32, 700 

33, 700 
33, 700 

99, 100 
32, 700 
31, 400 
98, 500 


5,860 

1,220 

130 

/50 
20 

1,010 
350 
340 
390 

80 

20 

280 

490 


57.2 




11.9 


4.05 


1.3 


^0.83 


.5 


* 0. 55 


.2 


• 14. 43 


9.9 


• 10. 67 


3.4 


• 10.0 


3.3 


• 11.7 


3.8 


'0.83 


.8 


* 0.55 


.2 


'8.96 


2.7 


i «4.9 


4.8 






♦73.3 ... 






1 10, 240 


100.0 











109 



TABLE 12.6.-TOTAL ENERGY IMPACT OF TRANSPORTATION, 1963 



Use 



Energy 
(lOis Btu) 



Percentage 
of total U.S. 
energy use h 



Direct fuel*.- 11.5 23.1 

Refining, and so on t> 2.3 4.6 

Highway construction '- .57 1.1 

Transportation equipment and maintenance d ' 2.33 4.7 

Transportation services*' 3.52 4.8 

Other expenses associated with private automobile: « 

Insurance.. .19 .4 

Accessories 0.06 .1 

Retail markup energy: 

Cars .26 .5 

Gasoline. 0.11 .2 

Total. i 39. 5 

» Bureau of Mines figure. See American Petroleum Institute (1971 ed., p. 443). 

*> 20 percent of direct fuel. See table 12.3. 

« I/O sector 11.04. 

<* Includes tires, maintenance, motor vehicles and parts, aircraft and parts, and other transportation equipment (I/O 
sectors 32.01, 75.00, 59.01-59.03, 60.01-60.04, 61.01-61.07). 

• Includes services such as air, water, rail, urban passenger and freight, pipeline transportation (I/O sectors 65.01-65.07). 

' Energy is obtained by multiplying total output (minus that sold to other transportation sectors), in dollars, times the 
appropriate energy coefficient from Herendeen (in press) (diminished by the energy used by industry and commerce for 
transportation). 

8 Various figures in Bureau of the Census (1971). There were 8.05X10 6 cars sold in 1963 at an average markup of $1,140 ; 
there were 69X10 6 cars on the road paying an average yearly insurance premium of $129; the average car owner bought 
$14 of accessories; 46X10 9 gallons of fuel were bought at a markup of approximately 9 cents per gallon. All these numbers 
were converted to energy using appropriate (diminished) I/O coefficients. 

h U.S. energy use=49.8xl0 15 Btu in 1963 (American Petroleum Institute, 1971 ed., p. 443). 

> Since undoubtedly some additional transportation expenses have been omitted, this is a lower bound for transporta- 
tion's total energy impact in 1963. 

TABLE 12.7.-ENERGY BREAKDOWN FOR AIR TRANSPORTATION (65.05), 1963 









PCINVB2 












Crude 














oil and 








T0TPCP3, 






gas ex- 


Refined 






total 


Row and description ' 


Coal 


traction 


petroleum 


Electricity 


Gas 


primary 


1. Food 


3.20 


0.30 


0.22 


2.28 


1.26 


1.36 


2. Construction 


.28 


.23 


.24 


.47 


.16 


.23 


3. Textiles 


.31 


.01 


.00 


.28 


.01 


.12 


4. Paper and lumber. 


7.33 


.23 


.10 


2.54 


1.69 


2.94 


5. Furniture.. 


.01 








.01 








6. Chemicals 


15.05 


.89 


.29 


11.03 


6.68 


6.07 


7. Rubber 


1.11 


.02 


.01 


.61 


.18 


.43 


8. Leather 


.02 


.00 











.01 


9. Stone, clay, glass 


.79 


.07 


.01 


.51 


.73 


.34 


10. Primary metals.... 


21.57 


.47 


.14 


8.62 


4.15 


8.48 


11. Fabricated metal.. 


.26 


.02 


.01 


.33 


.12 


.10 


12. Machinery 


.75 


.04 


.02 


.85 


.25 


.29 


13. Instruments 


1.14 


.05 


.02 


1.29 


.32 


.43 


14. Transportation equipment 


4.47 


.16 


.11 


5.60 


.80 


1.67 


15. Transportation services * 


11.95 


85.40 


92.88 


18.66 


32.51 


53.00 


16. Mining (metal, stone, fertilizer).. 


.89 


.06 


.02 


.80 


.48 


.36 


17. Coal mining 


.82 


.01 


.00 


.31 


.03 


.32 


18. Crude gas 


3.15 


2.96 


.11 


5.27 


4.07 


2.76 


19. Petroleum refining, and so on 


8.65 


6.89 


4.98 


11.19 


29.26 


6.98 


20. Electric utilities.. 


5.02 
.19 
.30 


.08 
.35 
.04 


.02 
.02 
.03 


8.40 
.06 
.16 


.78 

3.95 

.17 


1.74 


21. Gas utilities . 


.27 


22. Water 


.14 


23. Wholesale and retail trade 


2.09 


.39 


.31 


3.50 


1.36 


.93 


24. Finance insurance and business.. 


3.65 


.38 


.17 


6.11 


2.69 


1.45 


25. Medical 


.02 








.03 


.01 


.01 


26. Education 


.39 


.03 


.01 


.65 


.16 


.15 


27. Advertising 

28. Radio, TV, communications 


.04 


.02 


.01 


.07 


.09 


.02 


.96 


.11 


.07 


1.54 


.54 


.39 



» 2 and 3 see notes for table 12.2. 

* Almost all of use by row sector 15 is direct fuel use. 



2. The energy for all transportation. 

This is partly a definitional problem. Twenty-three sectors would be called 
transportation sectors: highway construction, motor vehicles plus equipment, 



110 

aircraft plus parts, other transportation equipment, transportation plus ware- 
housing, and auto repair plus services. As described by Hirst and Herendeen 
(1973), other categories apply some of their final demand expenses to support the 
automobile (for example, insurance). In addition, the refined petroleum sector 
fuels almost all transportation. Then an estimate for transportation's total energy 

52r',(prim)(x,-22^«)+ 2T',(prim)r I +E F) a2.1i) 

where T' (prim) is the total primary energy coefficient diminished by the refined 
petroleum used by industrial transportation, H is the set of 23 transportation 
sectors, and H f is the set of sectors which partly support the private car. Here, 
Ef — direct fuel used by transportation and that burned in the society to produce 
that fuel (from Table 12.3, this is roughly an additional 20%). 

Table 12.6 lists the results. In 1963, transportation's direct fuel use was 23% 
of the U.S. use, but the total energy impact amounted to 40%. This is probably a 
lower bound, as undoubtedly many transportation-associated activities have been 
neglected (such as industrial construction for parking, which is a capital expense) . 
I would estimate it as high as 45%. 

3. Energy breakdown for air travel, I/O sector 65.05. 

See Table 12.7. Unfortunately, this sums passenger and freight; we are currently 
involved in disaggregating it. About 53% of the total energy is fuel. This sort of 
analysis will have to be applied to transportation systems before the present 
hierarchy of transportation energy intensities (Hirst, 1972) (which is based on 
direct fuel use only) can be trusted completely. My opinion so far is that the 
ordering will not change, and that those of us who flew here did indeed use the 
most energy intensive mode available. 

REFERENCES 

American Petroleum Institute (1971 ed.). Petroleum Facts and Figures, 1971 
edition, New York. 

Berry, R., and Fels, M. (1972). "The Production and Consumption of Automo- 
biles," Report of the Illinois Institute for Environmental Quality, Springfield, 111. 
July. 

Bowes, A., and Church, C. (1970). FoodValues of Portions Commonly Used, 11th 
ed., Lippincott, Philadelphia. 

Bureau of the Census (1971). Statistical Abstract of the U.S., Department of 
Commerce, U.S. Government Printing Office, Washington, D.C. 

Bureau of Economic Analysis (1969). Input-Output Structure of the U.S. Econ- 
omy: 1963, Vols. 1 to 3, U.S. Department of Commerce, U.S. Government Print- 
ing Office, Washington, D.C. 

(1969, November). "Input-Output Structure of the U.S. Economy: 

1963," Survey of Current Business, U.S. Department of Commerce, Washington, 
D.C. 

(1971, January). "Personal Consumption Expenditures in the 1963 

Input-Output Study," Survey of Current Business, U.S. Department of Commerce, 
Washington, D.C. 

(1971, August). "Interindustry Transactions in New Structures and 

Equipment, 1963," Survey of Current Business, U.S. Department of Commerce, 
Washington, D.C. 

(1972). "Definitions and Conventions of the 1963 Input-Output Study," 

U.S. Department of Commerce, Washington, D.C. 

Herendeen, R. (1973). "An Energy Input-Output Matrix for the United States, 
1963: User's Guide," Document No. 69, Center for Advanced Computation, Uni- 
versity of Illinois, Urbana, 111., March. 

(in press). "The Energy Cost of Goods and Services," Environmental 

Report, ORNL-NSF Environmental Program, Oak Ridge National Laboratory, 
Oak Ridge, Tenn. 

Hirst, E. A. (1972). "Energy Consumption for Transportation in the U.S.," 
ORNL-NSF-Environmental Report No. 15, Oak Ridge National Laboratory, 
Oak Ridge, Tenn., March. 

, and Herendeen, R. (1973). "Total Energy Demand for Automobiles," 

Society of Automotive Engineers, presented at the International Automotive 
Engineering Congress, Detroit, Mich., January. 

Perelman, M. (1972). "Farming with Petroleum," Environment, 14, No. 8. 

Reardon, W. R. (1972). "An Input/Output Analysis of Energy Use Changes 
from 1947 to 1958 and 1958 to 1963," Battelle Memorial Institute, Richland, 
Wash., June. 



Ill 



ENERGY BUDGETS 



A series of articles exploring and reviewing the rapidly-expanding study of the 
energy costs of production processes, encompassing foods and agriculture, 
transport, materials, petrochemicals and products, and their importance for 
policy making. 



Energy costs : a review of 
methods 



P.F. Chapman 



Over the past two years there has been a growing realisation that 
the financial costs of materials and products do not provide an 
adequate description of the resources needed for their production. 
When there are no shortages of any inputs to the production 
system financial analysis provides a convenient decision-making 
framework. However, if one input does become scarce, then the 
implicit assumption of substitutability, inherent in financial 
systems analysis, leads to false conclusions. For a wide variety of 
reasons a number of investigators have focused their attention on 
the physical inputs, such as tons of steel and kWh of electricity, 
needed to make particular products. The forecasts of energy 
shortages coupled with the realisation that energy is an essential 
input to all production processes have concentrated attention on 
the energy inputs to, or energy cost of, various products. 

At the present time there are almost as many methods of 
evaluating the energy cost of a product as there are workers in the 
field. Where, by chance, the same product has been analysed by 
different methods the results often vary widely. The purpose of 
this review of methods is to explain the origin of the variations in 
results so that they can be interpreted and used correctly. To 
accomplish this it is necessary to examine the aims of various 
investigations since this explains many of the assumptions made. 
Finally, it is necesary to show how differing .assumptions and 
methods can account for the divergent results. 



K.E. Boulding, Economics as a Science 
(McGraw Hill, 1970), chapter 2 



The nature of the problem 

A modern industrial system, exemplified by countries such as the 
UK and the USA, is a complex interconnected system with many 
inputs and outputs. These highly developed systems are linked 
together, and to the so-called underdeveloped industrial systems, 
by flows of commodities in international trade. In some respects 
the total global system is a closed system. All that man's activities 
accomplish is a temporary change from a stock of raw material or 
flow of solar energy, into products such as automobiles and food 
which, in time, become discarded materials and dissipated energy. 
This view of the world - the 'spaceship Earth' concept^/- has 



ENERGY POLICY June 1974 



91 



112 



Energy costs: a review of methods 



Figure 1. Possible sub-systems associ- 
ated with the production of a loaf of 
bread. System 1 is denoted by dotted 
boundary, system 2 by dashed 
boundary, system 3 is the entire 
diagram. 



focused attention on the depletion of non-renewable stocks, 
particularly fossil fuels. Many analyses of energy costs aim to 
evaluate the quantity of fossil fuel energy required to produce a 
consumer product such as an automobile or a loaf of bread. 

The production of a consumer product in the UK requires 
inputs from all the production processes in the country and, 
through international trade, from all the production processes in 
the world. For example, a loaf of bread requires wheat which has 
to be milled, cooked and transported. Transport requires fuel and 
vehicles, for which steel, rubber, copper and energy for fabrication 
are necessary. Shops and bakeries need bricks, steel, cement, wood 
and glass; wheat production must have tractors, fertilisers, 
insecticides etc. It is clearly impossible to determine the propor- 
tion of all the production processes in the world needed to 
produce a loaf of bread, or any other single product. Any analysis 
must be based on a sub-system of the world, a sub-system for 
which all the inputs and outputs are known. The choice of 
sub-system is the first crucial step in evaluating an energy cost. 

Three simple sub-systems of the production of a loaf of bread 
are shown in Figure 1 . The First is confined to the bakery and the 



Energy cost of materials- 



Buildin, 
Plant 



Energy to deliver materials 



r ---- 



i 



Transport energy . 



Energy used in shop 



V, 



Transport energy - 



• 



92 



ENERGY POLICY June 1974 



113 



Energy costs: a review of methods 

energy cost per loaf is the energy delivered to the bakery divided 
by the number of loaves produced. The second sub-system 
includes the baker's shop. The total energy cost is: 



energy used at bakery transport energy energy used by shop 

loaves baked loaves delivered loaves sold 



The third sub-system is the entire diagram and includes eight 
energy inputs. As the sub-system is made larger the total energy 
cost continues to increase. However, in a finite time it is not 
possible to take into account all the production processes in the 
world. A more feasible objective is to follow each network of 
inputs back from the final product until it is found that the 
addition of the next input makes an acceptably small difference to 
the total energy cost. 

The choice of sub-system is one type of problem in evaluating 
energy costs. Another is associated with the types of energy 
included in the analysis and how these different energies are added 
together. The largest global source of energy is solar energy, yet 
this is usually excluded from energy costs. The production and 
delivery of fossil fuels involves energy consumption which may or 
may not be incorporated into the energy analysis. Producing 
secondary energy supplies, such as electricity, town gas and coke, 
wastes some of the energy available in primary fuels. This 
inefficiency may or may not be included. Most energy analyses 
ignore the energy input in the form of manpower or the calorific 
value of food. These difficulties are compounded by the various 
calorific values of different primary fuels and by the special role 
played by electricity in many industrial systems. 

A third type of problem which arises is in apportioning energy 
costs between different products. For example many chemical 
processes produce two or more products in a single plant from a 
single set of inputs. On what basis is the energy of the plant and 
inputs to be divided between the products produced? On a larger 
scale there is the problem of apportioning the energy costs of 
general services, such as roads, between many users. As with all the 
other problems outlined in this section there is no 'correct' 
solution. These are not questions of fact, but of setting up the 
most satisfactory conventions. Analyses based on different con- 
ventions will imply different procedures for dealing with the 
problems. 



Aims of energy studies 

The fact that there is no 'correct' way of apportioning energy 
costs or choosing a sub-system does not mean that these are 
arbitrary decisions. The methods adopted within a particular 
analysis should be consistent with the overall aims of the analysis. 
Thus the first step in assessing the techniques employed in a 
particular study is to establish the aims of that study. Although 
not usually stated explicitly the aims of most studies can be 
inferred from published reports. There appear to be four types of 
aim: 

ENERGY POLICY June 1974 93 



Energy costs: a review of methods 



114 



1. to analyse particular processes in detail so as to deduce an 
energy efficiency and hence make recommendations for con- 
serving energy; 

2. to analyse the consumption of energy on a large scale either to 
forecast energy demand or to point to policies which could 
reduce future demand; 

3. to analyse the energy consumption of basic technologies such as 
food production and mineral extraction so as to show some of 
the future consequences of technological trends or an energy 
shortage; 

4. to construct energy costs and examine energy flows so as to 
understand the thermodynamics of an industrial system. This 
type of long-range aim may be coupled to projects such as 
'world modelling' based on physical rather than monetary 
flows. 



I. Bousted, Journal of the Society of 
Dairy Technology (in press) 1974 

B.M. Hannon, Environment Vol 14, No 
2, 1972, p. 11 

A.B. Makhijani, and A.J. Lichtenberg, 
Environment, Vol 14, No 5, 1972, p. 10 

5 C.A. Berg, Science Vol 18, July 1973, p. 
128 

6 E. Hirst and J.C. Moyers, Science Vol 
179 No 4080 1973, p. 1299 

7 D. Pimental et al. Science Vol 182 
November 1973, p. 443 

G. Leach and M. Slesser, 'Energy equiv- 
alents of network inputs to food produc- 
ing processes' Strathclyde University, Glas- 
gow, 1973 

P.F. Chapman, Metals and Materials Feb. 
1974 

P.F. Chapman, 'Energy and world 
modelling', Seminar report. Open Univers- 
ity 1973. 

D.J. Wright, 'Calculating energy re- 
quirements of commodities from the in- 
put/output table'. Paper presented at Con- 
ference, Imperial College, London, July 
1973. 



This general classification of aims is neither exclusive nor inclusive. 
The aims are listed hierarchically so that a study under aim (1) 
could actually be part of an overall project with aims (2), (3) or 
(4). There are, no doubt, other aims not falling readily under any 
of these headings. 

Studies under aim (1) are often carried out by particular 
industries in order to make financial savings. The detailed study of 
particular processes requires data not normally published or 
generally available. An example of this type of study done outside 
industry is the examination of packaging. 2 ' 3 ' 20 By far the most 
popular type of study is associated with aim (2) since this 
corresponds most closely with 'energy policy' and the most 
obvious problems of the 'energy crisis'. Such studies 4 ,5 ' 6 are 
usually based upon published national statistics. Investigations 
with aim (3) are often aimed at areas, such as food production 7 ' 8 
and mineral resources, 9 where conventional economics is in 
conflict with the predictions of 'conservationists'. In these 
examples an energy approach throws a new light on complex 
problems. The investigations can be based on published data since 
great accuracy is not important and the conclusions are in terms of 
national or global averages. Only a few studies, 1 ° ' J * are associated 
with aim (4) on its own; however if aim (4) is the overall project 
aim then it has considerable influence on the assumptions made 
(as shown in a later section). 



Methods 

The implications of adopting and choosing different aims and 
conventions can best be illustrated by considering detailed 
examples. Before examining the results obtained by various 
authors it is necessary to outline the methods they have used. The 
fundamental principal of energy costing is that for a given industry 
or sub-system the total energy cost of all the inputs should equal 
the total energy cost of all the outputs. Thus if it requires 10 tons 
of steel (at 9940 kWh/ton steel) and 5 gallons of fuel oil (at 55 
kWh/gallon) to make 10 girders, the energy cost per girder is [(10 
x 9940) + (5 x 55)] + 10 = 9967-5 kWh. The methods used for 
calculating energy costs of products differ in their sources of data 



94 



ENERGY POLICY June 1974 



115 



Energy costs: a review of methods 

and techniques for deducing results. There are three types of 
method currently in use. 

Statistical analysis 

The supply of energy to various industries is available, for most 
industrial nations, in statistical publications such as the UK Report 
on the Census of Production, 1968. This information, coupled 
with data on industrial output, allows an estimate to be made of 
the energy cost per unit of output. For example the UK Digest of 
Energy Statistics gives the energy supplied to the iron and steel 
industry (1968) as 6871 x 10 6 therms. The output of crude steel 
(1968) is given as 25-86 x 10 6 tons {Iron and Steel Industry 
Annual Statistics). This gives a value of 265-7 therms/ton steel. 

This result is not a useful value for a number of reasons. The 
method has made no allowance for: 

• the energy cost in generating the electricity and coke consumed. 
The 6871 x 10 6 therms is the energy actually delivered to the 
industry; it requires about 8700 x 10 6 therms of primary fuel 
consumption; 

• energy sales by the iron and steel industry. Sales of gas and 
electricity in 1968 were 48 x 10 6 therms; 

• energy expenditures associated with the consumption of raw 
materials, the depreciation of plant or the delivery of materials 
and products. , 

However all these objections can be taken into account by digging 
a little deeper into the published statistics. In general this method 
can provide an order of magnitude estimate of the energy cost of 
products classified by industry. It cannot take into account all the 
subsidiary energy costs; nor can it distinguish in detail between 
different products of the same industry. Since this method relies 
upon national statistics the sub-system assumed is the nation. 

Input-output table analysis 

The input-output (I/O) table of a national economy is a square 
matrix, A, summarising the commodities necessary to make other 
commodities. Thus a single entry in the table, Ay, in the ith row 
and /th column, indicates the amount (measured in money) of 
commodity i required as a direct input, to produce £1.00 worth of 
commodity /. Thus all the inputs necessary to make £1.00 worth 
of commodity / are the items in the /th column of the square 
array. 

For a given set of outputs, denoted by a vector x, the direct 
inputs required, denoted by a vector y, can be found by 
multiplying x by the matrix (the I/O table) A : 

' y - A x 

To find the commodities, z, needed to produce the commodities 
y, the same procedure is adopted. Hence 

z = Ay=A (A x)=A 2 x 

Thus all the inputs, direct and indirect, required to produce the 
outputs x are A x + A 2 x + A*x + . . . This series can be summed 

ENERGY POLICY June 1974 95 



Energy costs: a review of methods 



116 



mathematically. The result of this analysis is thus a list of all the 
commodities required, within the nation covered by the I/O table, 
to produce a specified output. Clearly this method is taking a 
national sub-system and evaluating all the inputs within that 
system. 1 1_13 

There are some obvious disadvantages to this approach. Clearly 
the I/O table cannot be broken down into individual firms ; it has 
to deal with industries in groups. Another disadvantage is that the 
method deals with transactions in financial terms, not in terms of 
physical quantities. This can lead to errors if commodities are 
liable to large price fluctuations or if some purchasers can obtain 
special prices for the commodity. 

Process analysis 

Process analysis involves three stages. The first is to identify the 
network of processes which contribute to a final product, as 
illustrated in Figure 1. Next each process within the network has 
to be analysed in order to identify the inputs, in the form of 
equipment, materials and energy. Finally an energy value has to be 
assigned to each input. 

There are two clear problems with this method. The First is 
choosing an appropriate sub-system, the other is attaching energy 
values to particular inputs. This latter problem is crucial and can 
be illustrated by a simple example. The production of steel 
requires machines with a finite lifetime. Thus the energy cost of 
each machine must be averaged over all the steel processed by the 
machine. The machine will probably contain a great deal of steel 
and will have been made by other steel-containing machines. So to 
find the energy cost of the machine (necessary to find the energy 
of steel production) it is necessary to have an energy cost of steel! 
In practice this problem is solved by starting with an approximate 
energy cost of steel, 9 ' 10 (deduced by one of the methods above) 
and to use this to calculate the energy of the machine and hence a 
better value of the energy cost of steel. In most cases this feedback 
interaction only contributes a small percentage to the final energy 
cost estimate, so provided the final result is not wildly different 
from the starting value it is only necessary to go round this loop 
once. For industries which are strongly linked, ie where a 
significant fraction of each one's output goes into the other, the 
most direct way to solve the problem is to solve the simultaneous 
equations involved (shown in Figure 2). 



W.A. Reardon, 'An input/output 
analysis of energy use changes 1974-1975 
and 1958-1963' Battelle Northwest Labs, 
1971 

E. Hirst and R. Herendeen, Total 
Energy demand for automobiles'. Society 
of Automotive Engineers Inc., publ. 
730065. 1973. 



Results 

The following examples of results are intended to show the care 
required in interpreting bald figures. In all cases the aim is not to 
show that one result is 'wrong' and another 'right' but simply to 
underline the distinctions drawn in the previous three sections. 

Copper smelting 

This provides a convenient example of a detailed process analysis 
and the variation in results due to different choice of sub-sytem. 
There are various kinds of smelting furnace available but recently 
an electric furnace lias been recommended to the industry on the 



96 



ENERGY POLICY June 1974 



117 



Figure 2. Two industries, X and Y, 
supplying each other with raw 
materials. 















Energy costs: a review 


of methods 


i 


f > 






gyE r 




fxE, 








I) 








" 






Total mpuFf, ♦ jyf, 






Total inpur=f,<-r;r£ 






INDUSTRY X 








INDUSTRY Y 




1 




' I 


Output* units 
Energy cost E M fann 

1 




Outpu 


/ cost f_/jnit 


1 :. 


H-IxE, U-g)yE r 


For each industry total energy cost in eouals total energy cost out 


*£>F, *gr£, vE,*E-,*''e. 


Hence 




f , * g£ 7 


c *,♦*. 






' »0->9> 










' y 


|_/0| 







basis of better thermal efficiency. A detailed examination of the 
heat processes within an electric furnace 1 4 shows it has a thermal 
efficiency of 61% compared to fuel-heated furnaces with an 
efficiency of 27%. A comparison of the heat inputs required per 
ton of copper thus shows a factor of two in favour of electricity. 
Whether this represents a financial saving to a producer depends 
upon the relative prices (£ per unit of heat) of electricity and 
other fuels. However, as far as the industry is concerned, this is a 
significant energy saving. 1 5 

If the sub-system considered is enlarged to include the 
electricity supply industry and other inputs to the electric furnace 
the opposite conclusion results. The present efficiency of electric- 
ity generation in the UK is about 25% (see below) indicating that 
the supply of 1 kWh-electrical (kWhe) requires an input of 4 
kWh-thermal (kWhth). Thus the 2 to 1 ratio in favour of electric 
furnaces becomes almost a 2 to 1 ratio against. A detailed study of 
both smelting systems 16 shows that for a fuel-fired system the 
energy cost is about 5400 kWhth/ton copper and for an electric 
furnace about 8000 kWhth/ton copper. It is a disturbing con- 
clusion that in good faith an industry could improve its own 
thermal efficiency whilst increasing the national energy consump- 
tion. 



O. Barth, in Extractive metallurgy of 
Cu, Ni and Co ed P. Queneau (NY, 
Interscience, 1960) page 251. 

D.G. Treilhard, 'Copper state of the 
Art' Chemical Engineering, April 1973 

P.F. Chapman, 'The energy cost of 
producing copper and aluminium from 
primary ore', Report ERGOOI, Open 
University 1973 



Supply of electricity 

The < contradictory results obtained from two analyses of copper 
smelters hinged on the distinction between energy delivered (as 
electricity) and total energy input (to a nation). It is wortli 
exploring this topic further if only because different authors use 
different conversion factors and in any case the efficiency of 
electricity generation is likely to change with time. This example 



ENERGY POLICY June 1974 



97 



118 



Energy costs: a review of methods 



* A similar procedure is adopted for the 
electricity purchased from industry. The 
electricity is converted to 'tons of coal 
equivalent' and added to the 'coal input to 
the electricity supply industry'. 



will also show how the aims of a particular study can dramatically 
alter the answers obtained. . 

The Digest of Energy Statistics for the UK defines the primary 
input to the UK as coal, oil, gas, nuclear electricity and 
hydro-electricity. The latter inputs are converted to tons of 'coal 
equivalent according to the amount of coal needed to produce 
electricity at the efficiency of contemporary steam stations.'* This 
dubious procedure therefore introduces a theoretical loss into the 
energy supply system, as shown in Figure 3. The problem is 
thrown into focus by considering what energy cost should be 
attributed to one kilowatt-hour of electricity consumed. On the 
basis of the convention adopted by the Digest of Energy Statistics 
the total input is 22 784 x 10 6 therms, the output 5826 x 10 6 
therms giving an efficiency of 25-57%. This corresponds to an 
energy cost of 3-91 kWhth per kWhe. However it could be argued 
that the electricity output of nuclear and hydro-stations is the true 
input to the system. On this basis the total input is 20 979 x 10 6 
therms giving an efficiency of 27-77% so that the energy cost is 
3-6 kWhth per kWhe consumed. Alternatively, it could be argued 
that the inputs to the system are the fossil fuels (19 903 x 10 6 
therms), the heat generated at nuclear power stations (2360 x 10 6 
therms) and the hydro-electricity output (194 x 10 6 therms). 

This set of inputs gives an efficiency of 25-94% and an energy 
cost of 3-855 kWhth per kWhe. This last set of inputs is consistent 



Figure 3. Energy network of elec- 
tricity generation in the UK, 1968 
(source: UK Digest of Energy Stat- 
istics), (units: 10 6 therms) 





























Coal 
17 207 




O.I 
2687 




Gas 

9 




Nuclear 
2 360 




Hydro 
521 






\ 




! 


' 




i 


r 




1 


' 


1 








' 


' 












19 903 


























Conversion loss 
13 820 








Theoretical loss 
1 805 






















1 Electricity 
6 110 


1 Electricity 
|| 1076 
















' 


t 








t 








Total generated 
7 186 


























1360 
















1 Delivered to 

1 consumers 

5 826 







































98 



ENERGY POLICY June 1974 



119 



Energy costs: a review of methods 

with a project examining the heat release 1 7 involved in the supply 
and consumption of electricity. If the project was investigating 
maximum possible efficiencies then the primary input to nuclear 
power stations is the energy theoretically available in the 
fabricated fuel rods. This may be an order of magnitude larger 
than the heat extracted from the rods in the reactor. 

Apart from these differences there will also be alternatives in 
the ways in which the indirect energy consumption of power 
stations is taken into account. Indirect energy consumption is 
ignored in tables of energy statistics. In the case of nuclear power 
stations, how does one include the energy and materials consumed 
in processing and safeguarding nuclear wastes long after the power 
station has been taken out of commission? Table 1 summarises an 
analysis of the electricity supply industry based on the Census of 
Production 1968. ls It includes the purchases of electricity from 
industry, the consumption of materials and equipment, the energy 
needed to refine oil, mine and transport coal etc. (It does not 
include estimates of the energy costs of nuclear wastes or fuel 
reprocessing.) On this basis the energy cost of lkWhe is 4-19 
kWhth, corresponding to an efficiency of 23-84%.* 

In short this example shows that even when the sub-system is 
well defined there are alternative definitions of inputs which can 
dramatically alter the final values. 

Oil refining 

Another input to the energy supply industry is oil. Before any oil 
is consumed, either as fuel or chemical, it has to be extracted, 
transported and refined. The oil refinery is an easily identifiable 
sub-system; however there are differences both in what is counted 
as an input and how to partition the inputs between various 
outputs. A simplified flow diagram of the oil refinery system 



P.F. Chapman, New Scientist Vol 58, 
1973 page 408. 

18 P.F. Chapman, 'The energy cost of 
delivered energy, UK 1968', ERG 003, 
Open University 1973 

• In fact this value was obtained by 
solving the five simultaneous equations 
linking the major energy producing indus- 
tries. 



Table 1. The electricity supply industry 


Inputs 


10 6 kWhth 


New buildings (£8008 million) 


1 895 


Net plant etc (£462 million) 


16 170 


Net vehicles (£3-623 million) 


178 


Vehicle and machine spares 


766 


Iron and steel (13 500 tons) 


134 


Wire and cables 


1089 


Other materials 


985 


Nuclear fuels 


2 904 


Coal (75-54 million tons) 


542 015 


Coke (314 000 tons) 


2 944 


Oil products 


93 792 


Gas 


374 


Electricity purchased (from industry) 


3 479 


Heat input (nuclear power stations) 


96 113 


Electricity (generated in hydro-plant) 


3 600 


Total input 766 438 


Total electricity generated 


215 149 


Used in works, offices etc 


16 195 


Loss in distribution 


16 182 


Sold to final consumers 


182 772 


Overall efficiency = 


23-84% 


Hepce 1 kWhe 


4 195 kWhth 



ENERGY POLICY June 1974 



99 



120 



Energy costs: a review of methods 



* This is absurd because it indicates that 
'energy costs' could be reduced by using a 
chemical as a fuel. This is a good reason 
for insisting that the 'energy cost' of a 
product should be greater than or equal to 
its calorific value. See also ref. 18. 



connected to an 'organic chemicals industry' is shown in Figure 4. 
The problem is to decide how much energy to associate with an oil 
fuel and how much with an organic chemical. This is an 
interconnected system similar to that shown in Figure 2. This 
example has the added complication that the chemical feedstock is 
produced by the same plant as the oil fuel products. 

Two of the many possible conventions that could be adopted in 
approaching this problem are that: 

— since the crude oil is purchased as a primary fuel, all its calorific 
value is to be divided between the fuel products; 

— the calorific value of the crude oil is to be divided between all 
the refinery products in the ratio of the calorific values of the 
products. 

Following these rules the types of results obtained are: 

• All the inputs are set as costs against the fuel outputs. The 
chemical feedstocks therefore have no energy cost attributed to 
them.* On this scheme the efficiency of the oil refinery as a 
fuel processor is 82-4%. 8 Thus the energy cost incurred in 
consuming a gallon of refmed petrol is 53 kWhth, compared to 
an actual calorific output of 44 kWhth/gallon. Thus all 
industries consuming oil fuels are assigned greater energy costs 
than under the convention below; industries consuming 
chemical feedstocks have lower assigned energy costs. 

• The sum of all the energy inputs is distributed as energy costs 
over all the refinery products in proportion to the actual 
calorific values of the various products. (Thus the energy cost of 

















► £ 


\ 




1 
1 
1 
| 






F 


< 


> 


J^ 


1 

1 

_ J 






V 


t i 




' 


■ 




1 
l 

i 
i 

I 
I 
i 


Crude oil 
delivered 


_ 


I 




ol plant etc 



121 



Energy costs: a review of methods 

each product is its calorific value times a constant.) Note this 
increases the effective input to the oil refinery as compared to 
the convention above by virtue of the feedback of organic 
chemicals to the refinery. The efficiency of the refinery as a 
fuel processing plant is now 86%, ' 8 so the energy cost incurred 
in consuming a gallon of petrol is 51 kWhth. On this scheme 1 
tonne of plastics has an average energy cost of about 30 000 
kWhth; on the basis of the convention above the figure is 
. between 10 000 and 15 000 kWhth. 

The oil refinery is a sub-system which receives as an input the 
output of another sub-system, namely that shown as the extrac- 
tion and delivery of crude oil in Figure 4. The energy costs 
incurred within the extraction and delivery sub-system are 
equivalent to about 7% of the calorific value of the crude oil. 8 
However whether these energy costs are included as part of the 
input to the refinery depends upon the overall aim of the project. 
It would be consistent to ignore these energy costs in a project 
aimed at evaluating the energy costs of products to the UK since 
these costs are incurred outside the UK. If the project is associated 
with world energy costs then clearly these extraction and 
transport energy costs must be included as part of the input to the 
refinery. 



J.C Bravard, H.B. Flora, and C Portal, 
iergy expenditures associated with the 
sduction and recycling of metals' Oak 
dge. Nat. Lab. Report. ORNL-NSF- 
'-24. 1972. 

P.R. Atkins, Engineering and Mining 
urnal Vol 174, No 5, 1973, page 69 



Aluminium production 

This final example of energy analyses shows how an understanding 
of different methods and conventions enables sense to be made of 
apparently contradictory results. Several independent estimates of 
the energy cost to produce 1 tonne of aluminium have been 
published. A statistical analysis of US data yielded 67 200 
kWhth/ton; 4 an input-output analysis for the UK yielded 16 600 
kWhth/ton 11 and process analyses have yielded 91 000 kWhth/ 
ton, 9 64 200 kWhth/ton 19 and 64 300 kWhth/ton. 20 The esti- 
mates made by US authors refer to the energy cost per short ton; 
converting these results to metric tons gives 75 000 kWhth/tonne 
by statistical analysis 4 and 71 900 kWhth/tonne 1 9 and 72 000 
kWhth/tonne 20 by the process analysis methods. All these US 
studies assume a conversion efficiency for electricity of 33% 
whereas my own process analysis 9 assumes 23-8%. Converting my 
result to 33% electricity generation efficiency yields 72 000 
kWhth/tonne, so that all three process analyses are in excellent 
agreement. The statistical analysis result is slightly higher; however 
it refers to the energy cost per ton of rolled, not crude aluminium. 
The only remaining discrepancy is the value obtained by 
Wright' ' on the basis of an analysis of the UK input-output table. 
The 'aluminium' sector of this table does not distinguish between 
aluminium produced from primary ores and that produced from 
scrap material; nor does it specify whether the product is crude or 
semi-fabricated aluminium in the form of sheet, tube etc. Thus the 
energy cost deduced from the I/O table is an average energy given 
by 






SE S 



MERGY POLICY June 1974 



101 



Energy costs: a review of methods 



122 



where P is the quantity produced from primary ores, S is the 
secondary production and E p and E s are the respective energy 
costs. Assuming that E p equals 72 kWhth/kg, and E s equals 3 
kWhth/kg. This is close to the I/O table result; the remaining 
energy cost per ton of aluminium produced in the UK is about 14 
kWhtht/kg. This is close to the I/O table result; the remaining 
difference could be due to some energy used to fabricate sheets 
etc. 

Thus all these estimates of the energy cost of aluminium are 
self-consistent, they are all 'correct', but they are all based on 
different conventions. Clearly none of these results should be used 
in other studies until they have been converted to the conventions 
appropriate to other studies. 



P.F. Chapman, The energy costs of 
producing copper and aluminium from 
secondary sources'. Open University, Re- 
port. ERG 002, 1973. 



Conclusions 

The analysis of energy consumption can show ways of conserving 
energy and can highlight particular types of problem. However this 
is a new area of study and there is no uniformity between authors 
as to what conventions to use or what techniques are most 
appropriate. This means that results from energy studies should be 
carefully interpreted and only used when the following points are 
clear: 

• what sub-system of the world has been analysed; 

• which energy inputs to the system have been included in the 
analysis; 

• what calorific values are being used for primary fuels; 

• what efficiencies are being ascribed to the energy supply 
industries; 

• what conventions are being used to partition the energy costs 
within plants or industries. 

Neglect of any of these factors could produce misleading 
conclusions even in work based on accurate data. Moreover, any 
work leaving room for doubt as to its conventions in respect of 
these points must be interpreted with great caution. 



BIBLIOGRAPHY 



A list of some of the important literature, with brief abstracts 



Atkins, P.R., 'Recycling can cut energy 
demand dramatically,' Engineering and 
Mining Journal, May 1973, p 69 
Energy cost of an aluminium can based on 
operating plant data. Based on direct 
energy consumption only. 

Bravard, J.C ef al., 'Energy expenditures 
associated with the production and recycle 
of metals' 0RNL-NSF-EP-24, Nov. 1972 
(Available Oak Ridge National Labor- 
atory, Oak Ridge, Tennessee 37830, USA) 
Gives energy cost of magnesium, alum- 
inium, iron, copper and titanium from 
ores. Does not count energy cost of 
transport or of machinery, water etc. 
Recycling energy cost is simply the energy 



to remelt the metal. Note all results refer 
to energy per short ton. 

Bousted, I., 'Milk bottles and energy pro- 
blems' United Glass Magazine, 1973. 
Compares energy cost of glass and plastics 
milk bottles for delivering milk and 
indicates dependence on number of trips 
made. (Details of energy calculations avail- 
able in TS251;CU9, 'Milk bottle', available 
from Open University, Milton Keynes, 
Bucks, UK) 

Chapman, P.F. (a) 'The energy cost of 
producing copper and aluminium from 
primary sources', Metals and Materials 
Feb, 1974. 



Gives complete energy costs based on 
process analysis and includes grade de- 
pendence of energy costs. (Details of 
energy calculations available in Research 
Report ERG001;0pen University Energy 
Research Group, Open University, Milton 
Keynes, Bucks, UK) 

(b) 'Energy conservation and recycling 
copper and aluminium', Metals and 
Materials (in press) 1974 
Gives total energy cost of recycling includ- 
ing transportation and re-refining. Cal- 
culates potential energy cost saving which 
could result from increased recycling. (De- 
tails of calculations in Research Report 
ERG002, Open University.) 



102 



ENERGY POLICY June 1974 



123 



(c) 'The energy cost of delivered 
energy', UK 1968 Research report ERG 
003. Open University. 

Evaluates energy costs of coal, coke, gas, 
oil, and electricity from data in 1968 
census of production. 

(d) 'Energy and world modelling'. 
Seminar report May 1973. (Available from 
OU Energy Research Group, Open Univer- 
sity, Milton Keynes, Bucks) 

Describes the relationship between energy 
cost and heat release and between heat 
release and climatic change. Indicates pos- 
sible model based on energy flows. 

Grimmer, D.P. and Luszcynski, K., 'Lost 
Power', Environment Vol 14, No 3, 1972, 
p14 

Describes energy consumption of various 
transportation systems. Gives overall ef- 
ficiencies of different systems but does 
not include energy costs of machinery, 
plant etc. Shows that electric automobile 
20% efficient overall compared to gasoline 
auto efficiency of 10%. 

Hannon, B.M., (a) 'Bottles, cans, energy'. 
Environment Vol 14, No 2, 1972, p 1 1 
Describes complete energy cost analysis of 
different containers with and without re- 
cycling. 

(b) 'Aluminium cans'. Environment 
Vol 14, No 6, 1972, p 46 

Herendeen, R., 'The energy costs of goods 
and services', ORNL-NSF-EP-40 1972. 
(Availability: see Hirst) 
Explains the input-output method of 
energy analysis developed at Oak Ridge 
and gives results for some services. (Details 
of results for US 1963 available in 'An 
energy input-output matrix for the US 
1963' report, CAC 69; Centre for Ad- 
vanced Computation, University of 
Illinois, Urbana, Illinois) 

Hirst, E., (a) 'Energy consumption for 
transportation in the US' ORNL-NSF-EP- 
15, 1972. (Available from Oak Ridge, 
National Laboratory, Oak Ridge, Tennes- 
see). 

Energy costs devised from input-output 
analysis of US economy. Gives energy per 
ton-mile and energy per passenger-mile for 
different transport systems, (see also 
ORNL-NSF-EP-44, April 1973. 

(b) 'Energy use for food in the US' 
ORNL-NSF-EP-57 (Available from Oak 
Ridge National Laboratory, Oak Ridge, 
Tennessee). 

The energy use in food-related activi- 
ties computed from input-output table for 
the year 1963. Shows total energy used by 



food industries is 12% of total energy. On 
average 6-4 Btu of primary energy con- 
sumed in delivery of 1 Btu of food energy. 
This report is reviewed and brought up 
to date by E. Hirst, Science, Vol 184, 12 
April 1974, p 134. 

Hirst, E. and Herendeen, R. Total energy 
demand for automobile' Society of Auto- 
motive Engineers, Pamphlet 730065 (or 
'How much overall energy does an auto- 
mobile require', SAE Journal of Auto- 
motive Engineering Vol 80, No 7, 1972, p 
36 

Complete energy cost of automobiles 
based on input-output table analysis of US 
economy. Includes inputs from insurance, 
roads etc. Notable that fuel consumption 
only accounts for 60% of total energy 
cost. 

Hirst, E. and Movers, J.C., 'Efficiency of 
energy use in the US', Science Vol 179, 
No 4080, March 1973, p. 1 229 
Uses data on energy costs of transport (see 
Hirst and Hirst and Herendeen) and energy 
efficiency of space heating to outline 
strategy for energy conservation. 

Leach, G., 'The energy costs of food 
production' in The man-food equation ed 
A. Bourne, Academic Press, 1973. 
Wide-ranging review of energy in agri- 
culture including energy cost analysis of 
major UK crops. Discusses implication of 
energy expensive food with respect to 
developing world and with respect to the 
price of energy. 

Leach, G. and Slesser, M., 'Energy equiv- 
alent of network inputs to food producing 
processes' (Available: M. Slesser, Depart- 
ment of Pure and Applied Chemistry, 
University of Strathclyde, Glasgow, Scot- 
land.) 

Gives the energy costs of energy, ferti- 
lisers, transport etc based on published UK 
statistics. 

Makhijani, A.B. and Lichtenberg, A. J., 
'Energy and well-being', Environment Vol 
14, No 5 (1972), p 10 
Discusses relationship between total 
energy use and 'quality of life'. Gives 
complete breakdown of energy consump- 
tion in USA by 38 sectors. Includes energy 
cost of major materials. Energy costs 
deduced from annual statistics and are 
only approximate. 

Mackillop, A., 'Low Energy Housing', 

Ecologist Vol 2, No 12, (1972) 

Discusses energy cost of housing using 



Energy costs: a review of methods 

energy costs of materials derived from 
1968 Census of Production (UK). Energy 
costs are only direct fuel costs. 

Mortimer, N., 'Energy cost of transport, 
UK 1968' (Research Report ERG004. 
Feb. 1973. Available from Open Univer- 
sity Energy Research Group, Open Univer- 
sity, Milton Keynes, Bucks) 
Gives energy cost of road and rail trans- 
port in the UK which includes all indirect 
costs such as highways. 

Pimentel, D. et al, 'Food production and 
the energy crisis' Science Vol 182, 2nd 
Nov 1973, p 443 

Complete energy cost analysis of USA 
corn crops 1940-1970. Shows that whi|e 
yields yhave increased dramatically the 
ratio of energy out/energy in has remained 
substantially the same. 

Roberts, F., 'Energy consumption in the 
production of materials' Metals and Ma- 
terials, Feb. 1 974 

Gives a review of the energy cost of major 
materials now and those to be expected in 
30 years' time. 

Roberts, P., 'Models of the future'. Omega 
Vol 5, No 5, 1973, p 592 
Discusses the role of energy in an indus- 
trial system and in models of the future. 
Indicates the significance of energy cost in 
imposing constraints, especially the energy 
cost of energy. 

Slesser, M... 'Energy subsidy as a criterion 
in food policy planning'. Journal of the 
Science of Food and Agriculture, Vol 24, 
Nov. 1973 

Review of energy costs of a wide range of 
crops showing that they conform to a 
trend relating to yield and energy output. 

Smith, H., 'The cumulative energy require- 
ments of some final products of the 
chemical industries' Transactions (World 
Power Conf.) Vol 18 1969 
Gives energy cost of a number of import- 
ant chemicals and shows that the analysis 
of chemical industry is complex. Energy 
cost does not include energy in feedstocks 
nor all energy costs of plant, vehicles etc. 

Wright, D.J., 'Calculating energy require- 
ments of commodities from the input/out- 
put table' (Available: Systems Analysis 
Research Unit, Dept. of the Environment, 
Marsham Street, London) 
Outlines basis of method and gives results 
based on UK input-output table (70 
sectors). 



ENERGY POLICY June 1974 



103 



124 

I Fl AS 



ENERGY ANALYSIS 



Workshop report 
Report no 6 



125 



ENERGY ANALYSIS WORKSHOP ON METHODOLOGY AND CONVENTIONS 



25th- 30th AUGUST 197^ 



GULDSMEDSHYTTAN 
SWEDEN 



under the auspices of the 



INTERNATIONAL FEDERATION OF INSTITUTES FOR ADVANCED STUDY 

The Nobel House, Sturegatan Ik, 
Box 53^ S-102 ^6 
Stockholm, Sweden. 



and sponsored by 
The SWEDISH BOARD FOR TECHNICAL DEVELOPMENT and The GRANGES COMPANY 



126 



CONTENTS 

WORKSHOP SUMMARY 

NOTE BY RAPPORTEUR 

INITIAL SESSION 

FOREWORD 

LIST OF PARTICIPANTS 



SECTION 



ORIGIN OF ENERGY ANALYSIS 

TITLE OF THE FIELD 

THE USES OF ENERGY ANALYSIS 

UNIT OF ACCOUNT 

l».l. What is Energy? 

k.2. Resources: a free energy source 

U.3. Energy for transformations 

k.k. Available work 

U.5. Definition of term and unit of account 

U.6. Naturally occuring energy sources 

U.7. Some examples of GER 

k.f.l. Nuclear energy 

k.Q. Convention on energy rquirement of fuels 
It. 8.1. Gross energy requirement of fuel 
U.8.2. Energy requirement for energy 
U.8.3. Significance of these conventions 

k.9. Net Energy Requirement 

U.10. Process Energy Requirement 

U.ll. Energy requirement for energy 

U.12. Products as an energy source 

U.13. Summary of recommendations on terms of account 

U.lU. Units of account 

1+.15. Discarded terms 

WASTE 

SYSTEM BOUNDARY, DATA AND MAN-POWER 
6.1. System boundary 



3 
k 

5 
6 

10 

11 

13 

15 

IT 
17 

20 
22 
23 

2k 
2k 

25 

30 



31 
32 
32 
33 
33 
3k 
3k 

38 

kQ 

UU 



127 



2 

6.2. Data 1+3 
6.2.2. Recommendations 1+5 

6.3. Man-power 1+6 

6.3.1. Primitive system 

6.3.2. Agriculture in under-developed country 
6.3-3. Low-intensity agriculture 

6.3.1+. Factory process 

6.3.5- Recommendation for Energy Analysis 

7. PARTITIONING ENERGY REQUIREMENTS 57 

7.2. Possible conventions 57 

7.3. Recommendations 58 

8. PROCEDURE FOR ENERGY ANALYSIS 60 

8.2. Procedure 60 

8.3. Terminology 6l 
8.3.1. Abbrevations 

8.3-2. Energy Units 

8.3.3. Definitions 

8.1+. Flow-charting 62 

8.1+.1. Flow-chart symbols 

9. INPUT-OUTPUT ANALYSIS 67 

9.1. Description 67 

9.2. Theory 67 

9.3. Energy based input-output tables 70 
9.1+. Energy Requirement for Energy 71 

10. MARGINAL AND AVERAGE ENERGY REQUIREMENT lh 

10.2.. Marginal Energy Requirement 75 

10.3. Obtaining Actual Values 76 

11. ENERGY DISCOUNTING 78 

12. INTERFACE BETWEEN ENERGY AND ECONOMICS 79 

APPENDIX 1. 

Some comments on energy sources 80 

APPENDIX 2. 

The thermodynamics of available work 8l 

APPENDIX 3. 

Gross energy rquirement of nuclear fuel 88 



128 



IFIAS WORKSHOP ON ENERGY ANALYSIS 

GULDSMEDSHYTTAN 

SWEDEN 

AUGUST 26/30, 197^ 



WORKSHOP SUMMARY 

The workshop, comprising twenty persons from nine countries, 
considered in depth the methodology of Energy Analysis. It found 
that the form of the analysis depended upon the objectives of the analysis. 
In particular, where the main interest lay in the effect of energy use 
on resources, it recommended that analysis be carried out in Free Energy 
terms. Free Energy analysis also provided the most powerful tool for 
assessing economic potential and process waste. However it recognised 
the difficulty of obtaining Free Energy data for many systems. Where 
the energy source was a high intensity fuel, like oil, coal, nuclear, etc 
the workshop found that the error in using gross enthalpy did not exceed 
10*. 

The workshop made a number of recommendations as to procedure and 
proposed a number of conventions. 

In order to provide a coherent framework for reporting, it also 
recommended a number of descriptive terms. 



MALCOLM SLESSER 
WORKSHOP RAPPORTEUR 



129 



NOTE BY RAPPORTEUR 

This report of the energy analysis workshop represents the consensus 
view of the participants of the workshop. While there was considerable 
discussion on almost every item on the agenda, unanimity was achieved on 
almost all issues, and no unbridgeable differences emerged. 

The Rapporteur takes responsibility for the structure of the report 
and its precise wording. He has endeavoured to represent all views fairly. 
The report has been circulated to all participants prior to publication for 
comment and correction. 

THE REPORT 

No attribution is given to any views expressed during the workshop. 

planned 
Though the workshop consisted of ten/sessions spread over five days with 

several additional sessions, the report does not attempt to record the 

proceedings in the order they occurred, but rather in such a sequence as 

to make it easier for newcomers to the field to appreciate the issues involved 

in energy analysis and how calculations may be carried out in such a way as 

to be meaningful and comparable with those of other workers in the field. 



130 



INITIAL SESSION 

Monday August 26th: purpose of the workshop and introduction. 
Chairman: Dr Sam Nilsson, Executive Secretary, IFIAS. 

Dr Nilsson opened the proceedings of the workshop promptly at 9:00 a.m. 
by welcoming the twenty participants from nine countries. He outlined the 
purpose of IFIAS, since few of those present had had previous contact with 
it. IFIAS was established in 1972 under the auspices of the Nobel and 
Rockefeller Foundations, as a potentially significant new instrument 
for truly transdisciplinary and transnational efforts among the 
physical, biological and social sciences and humanities. In November, 
1973, IFIAS called a meeting of an advisory panel at the Niels Bohr 
Institute, Copenhagen to consider what, if any, were the appropriate tasks 
it could undertake in relation to energy. Two projects in the area of 
energy and quality of life were initiated. A paper presented by 
Amory Lovins of Friends of the Earth pointed out that a useful method of 
assessing energy matters was to calculate the lenergy content of goods and 
services. The advisory panel decided to convene a workshop to establish 
the ground rules for such calculations. Dr Malcolm Slesser of the University 
of Strathclyde, Scotland, was appointed to survey the field and invite an 
appropriate panel to the workshop. 

Before asking each participant to introduce* himslef , and his reasons 
for interest in this field, Dr Nilsson thanked the Granges Company of Sweden 
who had made available to the workshop their fine country house at 
Guldsmedshyttan, 200 km from Stockholm. 

Dr Nilsson emphasised that the purpose of the workshop was to establish 
a methodology. Every participant had, of course, objectives beyond 
methodology. These, he asked, should be subordinated to the primary 
purpose of the workshop. It was IFIAS' hope that within a year a larger 
conference would be held to discuss the applications of energy analysis to 
problems of policy, economics and quality of life. 



131 



6 
FOREWORD 

Energy may become a real limiting factor in many vital sectors of 
society because the resources are limited and because the extravagant use 
of energy may have unbearable consequences. 

The continuing energy crisis is leading to a new awareness that we 
need to know how much energy is used in producing goods and services. It 
might be a loaf of bread, starting with seed, grain, petroleum to make 
fertilizer and similar primary inputs and then deliver it to the supermarket. 
It might be a plastic or paper cup, an automobile, a bottle of beer, an 
aluminium- framed window - single or double glass, and many other things. 
So far we have measured such items in money terms - the price we must pay 
in the market. We will continue. to need to know money prices, but now we 
see a need to know not only the cost of energy, but how much energy is 
embodied in goods or services expressed as physical units. Such measurements 
include not only the direct energy use for fuel and to run machinery, 
but also the energy required to make available each of the input materials, 
tracing them back to their primary sources. This is what we call Energy 
Analysis. 

Four central questions may be asked about Energy Analysis: 

(1) When the market system works it reflects in the money price 
any increase in the cost of the energy content of goods or 
services. Why then do we need to know the energy content 
in physical units? 

(2) How difficult and expensive is it to determine the energy 
requirements of goods and services? Can energy analyses made 
by different people in different places be compared? Do 
they use the same methods for calculating energy content? 

(3) How would such information, if available, be used? What choices 
or decisions might be made which are different or better than 
such choices or decisions based on money cost alone? 



132 



(4) Conversely, if we do not have the information from energy 
analysis what mistakes of choice or policy may be made? 

As to the first question - why not rely on money prices - the 
following may be said. Correct functioning of the market system requires 
perfect information available to buyers and sellers and perfect competition - 
without the distortions and delays arising from government interventions 
through taxes, regulations, quotas, etc. and without concerted actions 
by sellers which modify the behaviour of "the free market". For example, 
trebling the price of oil has not trebled the price of gasoline because 
much of the retail price of gasoline is already tax. On the other hand, 
the price of fuel oil for heating, which usually bears little or no tax, has 
increased almost as much as crude oil and with relatively short delay. Sometimes 
governments intervene in the price system for political reasons - either to 
keep prices down or to move them above the level they might reach in a 
"free market". 

It is not clear how well current economic theory will guide us to wise 
choices as we encounter scarcities and discontinuities of supply. Behaviour 
of the market system provides little guidance in decisions on resource allocation 
for research and development, especially because it may take years or even 
decades before the research and development can have an impact. Because 
Energy Analysis is value free it can provide a valuable additional information 
for decision making. 

As to the second question - how costly and difficult is energy analysis - 
it appears neither costly or difficult. Much of the data needed exists 
already in national statistics, in industrial process data and in various 
other places. What has been needed most urgently is a set of rules and 
guidelines for calculating energy inputs and transformations, for common 
terminology and symbols, for units of measure, and for definition of the energy 
analysis framework. Without such commonly used measures and criteria no two 



133 



8 



analyses will agree. The purpose of this I.F.I. A.S. workshop has been to 
bring together some of the principal practitioners of this new activity in order 
that they might "hammer out" a set of proposed definitions and guidelines for 
energy analysis. This report of the I.F.I. A.S. workshop presents such a 
set of guidelines for use by those wanting to do energy analysis and for 
comment and criticisms which can be incorporated in a next edition of such 
Energy Analysis Guidelines. 

As to questions 3 and k - how to use the results energy analysis and 
what mistakes might be avoided by such use - the following examples may 
illustrate the utility of this new tool. The money costs of making industrial 
products vary substantially between different producers - in the same country 
or in different countries. Customary money cost accounting reveals the basis 
for some of these differences but not others. Energy analysis, based on 
physical units instead of money units, can reveal the specific energy inputs 
at every step in the process. Using the units, measures and symbols recommended 
by this I.F.I. A.S. workshop it is readily possible to examine the inputs at 
each step, as well as the total process, to probe into reasons for differences, 
and thereby to identify actions which can diminish the energy intensity of 
the process and reduce the energy use per unit of end product. Or, in another 
case, one may compare the anticipated energy saving of a development - say house 
insulation - against the energy investment to create the development. 

The customary engineering analysis of fuel savings and the financial 
analysis of return on investment will provide iulp6rtant data for comparing 
two or more alternatives. , La- -addition, : «the energy content of the materials 
and equipment embodied including capital in the alternative should also be 
added in and the results compared. Other things being equal, the less energy- 
intensive alternative should be preferred. Energy analysis is necessary to 
determine the energy required for each alternative. 

Not only can processes of manufacture be compared, but the energy 
required to make each unit of output can be seen. The energy impact of use 
of recycled materials can also be assessed. A company considering which of 



134 



9 
two or more substitutable items is to be produced would do well to compare the 
energy requirements of each in addition to the energy investment called. 
Governments should consider the total energy implications of their decisions. 
If rising energy scarcity is likely to characterize the future, the producer 
should weigh such choices and it is likely that governments will consider 
whether they should intervene by inducements or deterrents to influence the 
choice. 

Energy analysis can furnish new insights that can be utilized to develop 
quantitative parameters for use in private and public decision-making regarding 
energy use. In addition to the usual economic criteria, consideration of 
the implications of energy analyses can lead to more effective policy formation 
with reduced uncertainty. 



Carroll L. Wilson 

Massachusetts Institute of Technology, 
Cambridge. 

December 137k. 



135 



10 



LIST OF PARTICIPANTS 



Dr Frank Alessio, 

Director, 

Energy Modelling Programme, 

Electrical Power Research Institute 

3^12 Hillview Avenue, 

P.O. Box 10412, 

PALO ALTO, California 9703, U.S.A. 

Professor R. Stephen Berry, 

University of Chicago, 

Department of Chemistry, 

5735 S. Ellis Avenue, 

Chicago, Illinois 60637 > U.S.A~_ 

Dr Peter Chapman, 

Open University, 

Walton Hall, Walton, 

BLETCHLEY, Buckinghamshire, U.K. 



Dr Rurik Krymm, 
I.A.E.A., 
VIENNA, Austria. 

Professor Thomas Long, 

University of Chicago, 

Department of Chemistry, 

208 Kent Laboratory, 

5727 S. Ellis Avenue, 

CHICAGO, Illinois 60637, U.S.A. 

Dr Caesar Marchetti, 

I.I.A.S.A., 

2361 LAXEMBURG, Austria. 

Dr William Mebane, 
Montedison, 
MILAN, Italy. 



Dr Leslie Cook, 

Planning & Programme Analysis, 

EXXON, 

P.O. Box k5, 

LINDEN, N.Y. O7036, U.S.A. 



Dr Sam Nilsson, 
I.F.I.A.S. , 
The Nobel House, 
Sturegatan lk, 
STOCKHOLM. 



Dr Paul Craig, 

National Science Foundation, 

WASHINGTON, D.C. 20550, U.S.A. 

Professor Bent Elbek, 
Niels Bohr Institute, 
Blegdamsvej 17, 
2100 COPENHAGEN, Denmark. 

Dr Bertil Eneroth, 
Skandinaviska-Enskilda Banken, 
STOCKHOLM, Sweden. 

Dr Robert Herendeen, 
Centre for Advanced Computation, 
University of Champaign-Urbana, 
CHAMPAIGN- URBAN A, Illinois 6l801, U.S.A. 

Professor Takao Hoshi, 
Seikei University, 
TOKYO, Japan. 

Professor Ingvar Jung, 

Royal Institute of Technology, 

STOCKHOLM, Sweden. 



Dr Peter Roberts, 

Systems Analysis Research Unit, 

Department of the Environment, 

2 Marsham Street, 

LONDON, SW1P 3EB, U.K. 

Mr Joel Schatz, 

Office of the Governor, 

SALEM, Oregon 97310, U.S.A. 



Dr Malcolm Slesser, 

Director, Energy Analysis Unit, 

University of Strathclyde, 

100 Montrose Street, 

GLASGOW Gl 1X1, Scotland 

Dr Gerald Stanhill, 

Agricultural Research Organization, 

The Volcani Centre, 

P.O. Box 6, 

BET DAGAN, Israel. 

Professor Carroll L. Wilson, 
Massachusetts Institute of Technology, 
CAMBRIDGE, Massachusetts 02139, U.S.A. 



I.F.I.A.S. Secretariat: 



Eva Bengtsson 
Irene Holmgren 



136 



1L 

SECTION 1 



THE ORIGINS OF ENEEGY ANALYSIS 



All process technologists, whether chemical engineers or otherwise, 
are trained to consider energy economy in the design of processes. In 
high energy-use industries like electricity generation or ammonia synthesis, 
the direct energy costs of production have been high enough to encourage 
energy thrift through good design. Nevertheless it was rare to look at the 
energy requirements of the process beyond to immediate system boundary - 
usually the factory fence. The inputs to the process were thought of in 
tons, and their economic significance was assessed in terms of money, a 
commodity, as Gloyne (l) puts it that is 'best considered as a veil draped 
over the physical asset'. Money has not satisfied everyone as the only 
means for explaining the economic system. Soddy, the Nobel Laureate, 
virtually lost his reputation as a chemist in his efforts to argue that 
energy was the driving force of the economy. The inclusion of energy in 
physical rather than money terms has been toyed with by several people from 
G. N. Lewis and Ehrenfest to Hollis Chenery (12). Nevertheless it has 
never aroused great interest, for the fact is that until the step change in 
the price of oil in November, 1973» energy represented a mere k-6% of the 
cost of the inputs to an industrial economy. 

However, within the last few years an unease has been manifest, and 
some people have started looking at the amount of energy utilised in making 
goods and services, analysing processes all the way from ores in the ground 
to the finished article. This has led to the idea that energy is sequestered, 
or embodied in the good or service, and that one can talk about the energy 
requirement as so many MJ/^cg just as one might quote its cost as so many 
dollars per unit. 

The origins of energy analysis, are as diverse as the names given to it. 
However, the first papers with actual numbers in them only started appearing 
in 1971 (3,4,5,6,7,8,9,10), together with a book which provided a stimulus for 



137 



12 

many who were later drawn into the field. That book was ENVIRONMENT, POWER 
AND SOCIETY, by H. T. Odum (2). 

To several workers it was becoming apparent that a key parameter in 
the world economy was the energy requirement of energy. The publication 
of Berry and Fels' paper brought several workers into touch with each other. 
Leach and Slesser, for example, find they were both looking at agriculture, 
decided to pool their experience and jointly produced figures for the inputs 
to agriculture (ll). Even as this document was being printed some figures 
were found to be suspect, and they found that as their insight grew, an 
almost weekly revision was called for. Yet, though they were by now using 
identical energy equivalents of inputs they had a number of conceptual 
differences. For example what should one do about the energy to support 
the farm workers? 

Was there one unique truth, or merely a set of useful conventions? 
This was the state of the art in mid 197^- 



68-391 O - 76 - 10 



138 



13 
SECTION 2 



TITLE OF THE FIELD 



The workshop participants agreed that the purpose of energy analysis 
was to establish how much energy is required to make or provide a good or 
service. 

However, depending upon the objective of the analysis, one chose 
different conventions, and such conventions were by no means universally 
agreed upon. To establish appropriate conventions was a prime task of 
the workshop. For example, if the study in hand was to compare one aluminium 
process with another, the relevant energy content was that actually necessary 
to make the process function. There was no need to consider the energy locked 
up in the final aluminium product. In contrast if one was comparing the 
energy to make a glass bottle as opposed to a plastic one, one could reasonably 
argue that it was necessary to consider the energy sequestered in the plastic 
bottle itself because the input feed-stock has an alternative use as a fuel. 
What was needed was a descriptive title embracing the whole field from input 
calculations to subsequent aimlfkis. 

The literature contains several terms for energy analysis, reflecting 
the objectives of their sponsors. Ergonomics would have been an appropriate 
term, but this word already enjoyed a different connotation. It was felt 
that the word 'cost' which appears in several publications was misleading, 
since energy analyses were made in terms of a physical amount of energy, never 
in terms of money. The workshop examined and in turn dismissed the following: 
energy content analysis, net energy analysis, energy cost evaluation, energy 
systems analysis, energy value analysis, energy requirement, energy requirement, 
and others. It considered very carefully the proposal to call the field 



139 



"Energy Accounting", a word that had already gained a certain currency. 
Indeed the workshop, when convened, was called an energy accounting work- 
shop. Nevertheless the word is known to be frequently misconstrued. 
Moreover participants felt that the process of accounting for the energy 
used and/or sequestered in a given activity had as its ultimate objective 
an analytic purpose. Such analysis might refer to the activity under 
study, or might represent one bit of input data to a wider study. 

A proposal to refer to the field as enerconomics was closely examined. 
By a significant majority the participants were in favour of the term 
Energy Analysis. 

ENERGY ANALYSIS IS DEFINED AS THE DETERMINATION OF THE ENERGY SEQUESTERED 
IN THE PROCESS OF MAKING A GOOD OR SERVICE WITHIN THE FRAMEWORK OF AN AGREED 
SET OF CONVENTIONS OR APPLYING THE INFORMATION SO OBTAINED. The major part 
of this report is concerned with these conventions, and how they were logically 
arrived at. As a prelude it may be useful to consider some of the uses of 
energy analysis, as viewed by the participants of the workshop. 



140 



15 
SECTION 3 

THE USES OF ENERGY ANALYSIS 

The workshop felt that Energy Analysis could operate at three levels: 

1. To provide a marginal cost parameter leading to a cost (money) function. 
For example, which of several energy intensive processes, is the least 
energy intensive. 

2. As a means of identifying the constraints and boundary conditions 
of a system. 

For example, by means of thermodynamic calculations one may establish 
the theoretical energy requirements for a process, and compare them 
with those of present technology. This gave one a feel for the extent 
to which a given technology could be developed - a limit not identifiable 
by economic methods. 

3. As a means of injecting physical variables into economic theory. 

For example, Chenery (12), Long (13) i and others have considered adding 
economic 'production function. In practice, 
it was noted, we were becoming proficient in utilising the first level, 
and had barely made any progress at the third level. 

It was the view of most participants that the price system would 
remain the normal way to effect transactions and make decisions. Yet it was 
agreed the price system did not always embody sufficient information to 
make decisions, or to make them in adequate time. Energy represented a 
more sensitive signal of impending change, and identify those products where, for 
example, change of price of energy would most swiftly result in change in 
cost. 

The following is a list of those ways in which it was felt energy 
analysis could make a useful contribution. 



141 



16 
permits the internalisation of externalities 
can be a more sensitive indicator then money 

identifies where changes in the pattern of energy supply will impact 
most "speedily" upon the economy 
is a means of understanding price changes 
provides an attractive tool for allocation decisions 
ascertains the energy required to make goods and services 
provides a basis for estimating energy resource use 

offers a means of estimating the effect of technological changes and 
the potential for energy conservation 

provides a means of estimating the very important value of the energy 
requirements for energy for any given good or service far into the 
future 
adds a new primary variable for dynamic modelling of the economy 



142 

17 
SECTION h UNIT OF ACCOUNT 

k.l What is Energy ? 

Like money, the unit of energy account is not as simple a 
concept as it might at first sight appear. We can, of course, speak 
of some physically definable energy unit like the joule or kilowatt- 
hour, but we are still left asking "a joule of what?" For many workers 
it has been enough to define "what" as the calorific value of a given 
fuel, that is, the enthalpy of combustion of the fuel. Enthalpy is 
the thermodynamicist's description of heat, but to analyse the production 
of a" good or service in terms of heat can in certain circumstances be 
dangerously misleading. There is almost certainly more heat in the 
Atlantic Ocean than in the enthalpy of combustion of the oil in the 
middle east. It is some other quality than heat that makes oil an 
attractive fuel for driving the economic process. That quality is 
"free energy". ,»«j 

As is clearly enunciated in the first law of thermodynamics, 
Energy is always conserved. It is never lost. Its quality is merely 
degraded. The driving force for transformations, however, is not heat 
but thermodynamic potential, Free Energy. Free Energy is irrecoverable, 
and diminishes every time a fuel is burnt or a nucleus fissions. 

The workshop considered what should be the relevant unit of account, 
examining Enthalpy, Free Energy and Availability. It concluded that 
there was no unique input, and that it was necessary to adopt a 
convention. It was agreed, however, that the unit that best expressed 
the objectives of Energy Analysis is Free Energy (G) rather than enthalpy (H) 
but noted that for most intensive fuels (high free energy potential/unit mass) 
(oil, coal, etc.) the error in taking Enthalpy rather than Free Energy 
was of the order of 10# (see Table k.l). For many processes it is 
rather difficult to compute the Free Energy changes. For example, 
when the reactants and products are single compounds, of known chemical 



143 



18 
structure, as in the polymerisation of ethylene and polyethylene, or 

in the reduction of iron ore to steel there is sufficient thermodynamic 

data available to calculate actual and ideal Free Energy changes. 

Where the reactants and products are diverse, as in making ferro-concrete 

from aggregate, steel, and cement the data is simply not available for 

all the materials involved. Where the system is extremely complex, 

as in the growth of plants, or in human nutrition, we simply do not 

know enough to estimate the entropic contribution to free energy to 

better than one order of magnitude, even when we can measure the 

energy contribution fairly accurately. Before stating the conventions 

recommended it may be useful to amplify the thermodynamic argument. 



144 











•"> 












* 












3 


1 — 1 










•H 











<H 





H 








o 





XI 

ad 








£ 





p 








H 


H 


« 








(d 


x> 


X! 








Xi 





P 








■P 


■P 








• 


C 




<H 






-* 















t-- 




A 








CT\ 


a 


P 


p 













h 











«H 









a 


b 


o 


a 






o 




p 


u 






-p 




h 


© 






« 


p 





> 







o 


o 


a 







H 


s 


3 




H 







•H 


■o 


h 




X! 
p 


e 


£ 


o 
u 





4) 

XI 


■H 






o. 


a 


P 


* 


oo 


^ 




3 






3 


>■> 


o 




3 


T3 


o 


•P 


CM 





■H 





•H 


• 


K 


A 




P 


h 


■H 




•p 








ed 


O 


-o 




P 


■H 


> 


o 


•H 


3 


•H 


(J 




w 


3 


•H 


i 





Cm 




o 1 




(0 


o 


H 


Tt 







<n 




at 


H 





u 


«< 


c 







•rl 







o 


•H 





t)0 


xi 


oo 


•H 


00 




h 


p 





P 


* 


Ofl 








•H 


3 


£ 


3 

•H 


& 


s 


Jh 


XI 




■o 




•H 


0) 


E 


o 


8 

o 


H 


•H 


M 


3 








•H 


^ 


m 


> 






h 






■H 


t- 




• 


u 


• 


60 






1 


•H 


en 




<r 







r*- 


in 










cm 


•H 


3 




£ 


(3 
•H 


•o 


M 


pq 




■d 




s 


h 


< 




3 


a 


o 


H 




p 


o 




* 






(0 


•H 













P 


(4 









c 





3 


H 






o 


3 





XI 






ri 


X> 












■p 


S 





H 






cd 





U 


•H 






t 





a 



> 






• 


u 


h 


«< 






• 


o 











(3 


<H 


X) 








o 












o 





*~ 


H 







to 





P 




^ 






u 
















3 


^-N 


<H 






a 


o 


A 


o 






w 


•H 


o 










•P 


•H 














•P 


H 






M 


H 





O 






« 


3 


3 


E 






o 


o 


X> 








u 


H 


E 


U 






3 





O 












O 


o 


a 






CO 









"0 


It 2 


<y-P k 


■H O CO 


J o o 





3 





H 


►» ■* 


A X 


•P CM 


W O 





s 


A 00 


O X 


U en 


ft o 





§ no 


Xi X 


+> cm 


w o 





l«* 


4* O 





as ••.. 





■o 


fl n 


o K o 


XI o o 


h 3 


O 


o ac 


,-v 


3 M 


o o 


X» O 


U O 


O 


O -P 


*-*" 


3 





bo 


O CM 


h K 


T3 


S 


K 





g 





H 


fc 


h 





fl 



19 



r^ 


^fr 


m 


VO 


t>~ 


• 


^t 


CM 


r\ 


m 


tr\ 


I 


m 


00 


-#. 






ON 


o 


NO 




^ 


CM 


IfN 


-* 


r\ 


1 


*" 


*" 




^t 




**. 




0\ 


^t 


On 


On 




^> 


O 


»n 


CM 


CM 


I 


CM 






■4" 


en 




• 


• 


"* 


ON 


O 


>t 


ITN 


NO 




m 


-4 


NO 


■" 


*~ 


1 


^_ 


CM 


>R 






NO 


o 


cn 




ON 


»- 


0Q 


CO 


f*i i 


** v t: 


r>- 


oo 


-#- 


• 


• 


CM 


CM 


NO 




00 


m 


ON 


CM 


CM 




VO 


■4- 


* 






CM 


r\ 


-* 




ON 


ON 


O 


en 


cn 


+ 



00 


^ 


• 


00 


en 




en 


1 


CM 





<M 

> <H 
-H 

T3 



B0S 

H 

<M X A -P •< -P< 

O 3 3 

< H 

-P -H Xi OB 

1 -P (4 O 

> -H h 

W «< > ft <H 



145 

20 
k.2 Resources: a free energy source . 

One can consider the world as a vast store of free energy, 
available as oil, coal, shale, gas, peat, ocean temperature gradients, 
geothermal heat, uranium, deuterium and so on. In addition there is 
a daily flux of solar energy, which since it comes from a very high 
temperature source can, using appropriate technology or bio-mechanisms, 
be directly used as Free Energy. Some of the stored fuels readily 
liberate their free energy by direct combustion. Some, like 
uranium or shale require prior processing. Though the inflow of 
Free Energy from the sun is approximately 15,000 times greater than our 
current global use of stored Free Energy, we have no reasonable technologies 
yet outside agriculture, for using that solar flux, and are therefore almost 
entirely dependent on stored Free Energy. Thus, where the object of 
the energy analysis is to demonstrate how much the global store of 
energy is being depleted in making a good or service, Free Energy (G) 
is the relevant physical attribute. It is related to the Enthalpy (H) by 

G = H - T S where T is the temperature at which the 
transformation is carried out and S is 
the entropy change as a result of the 
transformation. 

Berry and Fels (ibid) established G for making an automobile in 
in Detroit, and found it differed from H by approximately 10#. 

The value of using G as a criterion of resource base is that it 
permits the inclusion energy ^e^ourcesfbiisjl ofi ifa intensity, but 
vast quantity like the oceans by virtue of their temperature 
gradients. (See appendix 1) The workshop would like to record its 
view THAT WHERE ENERGY ANALYSIS IS CONCERNED WITH DEPLETION OF THE 
RESOURCE BASE ALL FIGURES SHOULD BE EXPRESSED IN TERMS OF FREE ENERGY. 
HOWEVER, RECOGNISING THAT IN MANY CASES IT IS IMPOSSIBLE TO COMPUTE 
THE FREE ENERGY CHANGES OF ACTUAL PROCESSES, IT IS SUFFICIENTLY ACCURATE 
IN THE CASE OF INTENSIVE FUELS TO EXPRESS FIGURES IN TERMS OF GROSS 
ENTHALPY, IF CLEARLY STATED AS SUCH IN THE REPORTED WORK. 



146 



21 
TABLE 4.2 

Effect of adopted convention in attributing heat of combustion 

to methane. 

Source: Haslam <fe Russell) Fuels and their combustion, 
McGraw-Hill, New York. 



Heat evolved at 15.6°C, products cooled 

to 15.6°C, water condensed 
Heat evolved at 1000°C with initial and 

final states at 1000°C 
Heat available at 1000°C with reactants at 

15.6°C, and products of combustion 

at 1000°C 
Heat available at 1000°C, using- air 

instead of 0„ 



ADOPTED CONVENTION 

Heat available at 0°C (273.15K), initial 

and final statesO C, one bar pressure 



Enthalpy of 
combustion fro* 
1 kg methane 

55.704 mJ/kg 



49.43 



42.79 



28.72 



55-63 



This is known in fuel technology as the gross 
enthalpy of combustion. 



147 



22 
k. J Energy for transformations 

When a transformation is effected, such as the conversion of 
iron ore to iron, or oil to a polyethylene bottle the initial and final 
states have a unique value for both Enthalpy content and Free Energy. 
To compute the theoretical amount of Free Energy required to effect 
the change one needs only to know the initial and final states, and 
compute the change in Free Energy (AG). It is the change in G • 
the final state which is of consequence. Similarly by letting a fuel 
undergo a spontaneous change Free Energy (AG) may be released. By 
far the commonest, and most readily carried out spontaneous change is that 
of combustion with air. The workshop recommended that the Free Energy and 
Enthalpy of a fuel be taken as the gross AG (or AH) obtained when the fuel is 
combusted with air at one bar pressure and C (273- 15*0 with the products 
of oxidation restored to this temperature. Table 4.1 compare this value 
of AH with other conventions. The implication of "gross" is that the 
latent heat of the water vapour produced in combustion is condensed. The 
reason for choosing the gross and not the net energy of combustion is that 
this more correctly measures the total energy release to the biosphere and 
can be used in assessing the effect of energy release on climate. Table k.2 
gives some idea of the difference in values of H depending on the convention 
adopted. G is linked to H of the fuel: 

G m H + TAS summing over reactants and products. 

We must now distinguish between the theoretical and actual 
amount of Free Energy needed to effect a transformation. Where we 
are in the position to compute the theoretical (that is, minimum) 
Free Energy to convert (say) iron ore to iron, we can then (in principle) 
compute the minimum amount of any fuel that would furnish this amount 
of Free Energy. Such a computation would apply to common reference 
states, and would imply infinitely slow, reversibly-effected processes. 



148 



23 
Since the supply of Free Energy must equal its consumption, one 

may calculate the minimum physical amount of fuel that would provide 

the necessary Free Energy under the given reference states if one 

knows the minimum Free Energy required to effect a given transformation 

(such as iron ore to iron). This sets the thermodynamic minimum 

beyond which no technology can carry further. But it is also of interest 

to know how much fuel would be theoretically consumed, since our interest 

in the fuel consumed is to know the physical energy resource. One of 

the objectives of Energy Analysis is to ascertain the minimum actual 

Free Energy required to effect real transformations. 

k.k Available Work 

The best assessment of the quality of a fuel is Available Work (A). 
Available Work is an engineering concept. It takes into account the 
fact that the amount of useful work that can be extracted from the 
combustion of a fuel depends not on the total amount of heat liberated 
by combustion, but its temperature and the temperature of the sink. 
The workshop did not recommend the use of Available Work (A) as a 
preferred criterion for Energy Analyses, though it is undoubtedly an 
excellent criterion of efficiency of energy use. 

Appendix 2 discusses the thermodynamics of Available Work (Availability). 
In essence the difference between taking Available Work and Free Energy 
is that in the former pressure and temperature refer to the surroundings, 
whereas in the case of Free Energy, they refer to the reference state. 



149 



2k 

k.$ Definition of term and unit of account 

The workshop recommended that the unit of account be defined as 
THE AMOUNT OF ENERGY SOURCE WHICH IS SEQUESTERED BY THE PROCESS OF MAKING A 
GOOD OR SERVICE AND THAT IT BE NAMED GROSS ENERGY REQUIREMENT (GER) in place 
of the many other useful but slightly ambiguous terms such as energy content 
or energy cost. The value of GER is the gross Enthalpy released at standard 
state (as already defined) of all the naturally occurring energy sources which 
must be consumed in order to make the good or service available. GFER is de- 
fined similarly, in which Free Energy replaces Enthalpy. Since this definition 
embraces all the energy sequestered, it is necessary to add two further 
definitions to deal with those situations where the analysts wish to examine 
only the energy consumed promoting a given process or where he wishes to 
discount the energy locked up in the product. The former is referred to as 
the Process Energy Requirement (PER) and is the same as that obtained from an 
engineering heat balance. The latter is referred to as the Net Energy 
Requirement, and is treated in section k.9. 

In order that this definition be fully understood it is necessary to 
define a naturally occurring energy source. 

The preferred unit of account, in harmony with general international 
agreement is the joule or powers of ten thereof. 

k.6 Naturally occurring energy sources 

*f.6.1 Resources: some energy sources are fuels, which release their energy 
by a process of chemical reaction, usually with air (coal, oil, gas, etc.) or 
by transformation (Uranium 235 etc). A fuel is a naturally occurring substance 
which can be made to yield a negative change of Free Energy (-AG) along or in 
combination with other naturally occurring substances. The finiteness of a 
fuel is implied by the tefm "energy resource. ' An energy resource has the 
potential capacity to provide a finite, if vast amount of units of GER. 
6.6.2 Flux Sources: sources like solar radiation are natural flux sources 
from which a flow of energy occurs over extended periods of time. The potential 
of an energy flux source should refer to the maximum average rate of supply of 
G expressed relative to standard state already defined. The word "resource" £iould 
not be used for a flux source. See appendix 3 for some comment on this definition. 



150 



25 

4.7 Some examples of GER 

The workshop was conscious of the ambiguities that even these 
conventions could create, and felt that they could only be clarified 
by a series of examples. 

Fig '+.1 depicts a hypothetical process for making a good, y, 
which requires feed stocks of natural gas, silicon dioxide and hydrogen. 
It also requires an input of direct energy to make the process function. 
Let us suppose that this direct energy is furnished by coal. 

Three system boundaries are shown. The inner one (-.-.-) is that 
most immediately apparent to the person concerned with operating the 
process. The necessary inputs are coal, silicon diaxide, natural gas 
and hydrogen. The energy requirement of the process might be calculated 
by considering, at standard reference state, the gross enthalpies 
of combustion of the hydrogen, natural gas and coal. However one 
may readily see from Fig. h.l that this is not really the true energy 
requirement of the process. Hydrogen is not a naturally available 
substance, and has to be made by a process further back in the system. 
Such a process consumes more energy than is available in the hydrogen 
product. There is an energy requirement for making hydrogen. 
For the sake of example, let us assume that the hydrogen is made from 
natural gas. Is natural gas a free and infinite good? It is not. 
It has to be won by the process of exploration, production platforms, 
pipeline investment and separation plants. In order to make the 
energy of natural gas available to the processes within the middle 
system boundary (----- ) some energy must be expended on the 
supply system. Similarly, Silicon dioxide rarely occurs as a pure 
substance. It has to be mined, purified and transported. There is 
a necessary supply system. Equally, coal has to be mined, prepared 
and transported. There is a coal supply system requiring energy. 
These supply systems take operational energy from various sources, 



151 



26 

each of which in turn requires a supply system and energy source. 
Gross Energy Requirement (GER) is the sum of all the energy sources 
that must be sequestered in order to make product, y, available. 
It therefore includes everything passing through the outer system 
boundary (----). 



152 



27 



k. 7.1 Nuclear Energy 

The above calculation in terms of conventional fossil fuels 
presents no great problem, once the conventions are applied. Supposing, 
however, that in Fig. ^.1 the coal energy used for driving the process 
was partly replaced by electricity, which could be made from coal 
or nuclear energy. 

Let us take a case where 1000 kwh 3.6, GJ of electrical energy are 
delivered to the process and compare the methods of assessing the 
Gross Energy Requirement for delivering that amount of electrical 
energy by a coal and a nuclear source. Fig. k.2 depicts the situation 
for coal-generated electricity. In order to generate the 1000 kwh (3.6 GJ) 
of electricity not only has a certain amount of coal to be burnt, 
the amount depending upon the efficiency of operation of, the generating 
station, but some output electricity will be used to operate electrical 
devices in the power station. In addition some electricity may be 
used, together with other fuels, to operate the coal supply system. 

A recent analysis of the fuel requirement for supplying energy 
is given in appendix k. If nuclear energy is ignored, there are no 
ambiguities, and the methodology essentially responds to the equation 

x. GER. = y. GER. 

inputs outputs 

This assumes all fossil fuels utilised in the process are irrecoverably 
consumed. There are no unburnt fossil fuels left after the combustion 
process is finished. This assumption is justified because the 
equilibrium constant for combustion with air of most fossil fuels is 
extremely large, of the order of 10 or greater, and the kinetics fast 
and it is possible with negligible error to assume complete combustion. 
Gross Energy Requirement of electrical energy from a fossil fuel power 
station may therefore be based on complete combustion. 



153 



28 

The situation with nuclear energy is different. No existing 
nuclear fission process goes to completion (i.e. not a fissionable nuclei 
fission). Thermal reactors commonly give little more than 1% burn-up 
of the nuclear fuel. How are we to treat the 99$ discarded? Fig. k.J> 
depicts the situation. Let us assume for argument's sake that the depletec 
fuel is recycled, but for every 1% burnt-up, 1% is rejected in the fission 
product wastes. 

The definition of Gross Energy Requirement is that it includes all 
the energy sources sequestered in the process of making the good, 
the good in this case being 1000 kwh of electricity. Strictly 
speaking therefore it is the sum of the energy released by fission 
of U necessary to make the electricity PLUS the energy potentially 
releasable (undeveloped) in the rejected fuel. This rejected fuel 

is highly radioactive, being mixed with fission products. On 

235 
present technology there is a limit to how much U can be recovered 

from the depleted fuel and recycled. 

It will help to have some numbers. Assume that the nuclear 

generating station operates at 30$ thermal efficiency, but that 

allowing for the electricity needs of the Uranium enrichment plant 

and other aspects of Uranium supply, the net efficiency (in thermal 

terms) is 25#. Since ^.386 x 1Q~ kg of ^ 5 U will fission to produce 

1000 kwht of heat at standard conditions, 1/.25 of this amount will 

be needed to generate 1000 kwh of electrical energy that is 17»5^ x 10 \g 

235 

U. (Appendix 3) If there was no energy value in the spent 

fuel, then the system efficiency would be 25$ in thermal terms or 

k units of input heat per unit of output heat, with three units 

wasted. 



154 



29 



However if, as logic demands, we include the undeveloped energy 

235 

in the U lost in the fission product stream, then for each unit of 

delivered energy we are furnishing 8 units of input energy, and wasting 
7 units. This is closer to the real thermal efficiency of a nuclear 
power station. One has only to imagine the case in a coal fired 
power station where half the calorific value of the coal is lost because 
of imperfect combustion. An Engineer who attempted to argue that the 
efficiency of the power station was still around 30# because in 
principle the energy in perfectly burnt coal could be recovered at 
some future time would have difficulty in making his point stick. 



The workshop considered these points at great length over several 
days, and through two sub-committees, and concluded the only course 
was to adopt one of two conventions, either of which must be explicitly 
stated. These conventions are now summarised. 



155 

30 

k.8 CONVENTION ON THE ENERGY REQUIREMENTS OF FUELS 

Convention ^.8.1 

Gross Energy Requirement of the fuel. (GFER or GER fuel). This 

value is the total amount of fuel input whether consumed, rejected, 

transformed, or stored in waste products required in order to produce 

a unit of delivered energy, all expressed in terms of gross enthalpy 

of combustion or heat release per unit fission of a nucleus at standard 

state. 

This value is of especial interest where the objective of the 
analysis is related to the depletion of energy sources. 

Convention ^.8.2 

Net Energy Requirement of fuel. (NER fuel). 

This value is the Gross Energy Requirement for fuel as defined in lf.8.1 
less the energy equivalent of any unconsumed energy, such as may be 
stored in waste products or other by-products. The workshop recommended 
that this convention be normally used when assessing the energy require- 
ment of a good or service. 



Jf.8.3 Significance of these conventions. 

The significance of these conventions may be gathered from the 
following examples. 

Coal as a fuel : The GER of fuel would make no allowance for energy 
imperfect combustion or unburnt material in the ashes. NER fuel 
would make an allowance for these two possibilities. Normally 
such amounts would be trivial, and the difference between GER fuel and 
ERE would be trivial. 

^U as a fuel : GER fuel is a measure of the total ^U utilised in 
the process of releasing the energy in Uranium bearing rocks. 



156 



31 
The NER fuel (or ERE) would give a credit for any 2 -^U sequestered in 
rejected fission products. The difference between GER fuel and ERE 
would be large and important. 
h.9 Net energy requirement 

The conventions k. 8.1 and 4.8.2 help us to deal with the thorny 
problem of using fuels as feed-stocks. Two examples may suffice. 
Petrochemicals 

Fig. k.k depicts a process for making a polythene bottle. Under the 
definition of Gross Energy Requirement, the energy requirement for the 
bottle comprises the energy required to drive each stage of the overall 
process plus the energy to make the energy available plus the energy in 
hydrocarbons sequestered in the final product. This definition is 
useful for two situations. The first is where one wishes to analyse the 
manufacture of a polyethylene bottle in terms of depletion of energy sources. 
The second is where there is no possibility of using the energy (that is 
to say gross enthalpy of combustion) of the plastic bottle at some future time. 

It was for this kind of situation that the workshop recommended the 
definition of a unit called NET ENERGY REQUIREMENT which is the Gross Energy 
Requirement less the gross enthalpy of combustion (referred to the standard 
state) of the products of the process under study. As will be seen in 
section 4.9, it is important to include in the NER any energy required to 
carry out the process of releasing the energy from the product. In the 
practical case of polyethylene bottles, this is about 0.4 GJ/tonne. The 
bottle may be recycled or burnt. At the other extreme is uranium contained 
in the glaze on a dinner plate; the undeveloped energy in the uranium 
is large in relation to the PER of the plate, yet actually acquiring that 
energy may be impossible and expensive. See section 4.10. 

Where the product has a use as a potential fuel, GER, PER and 
NER all provide valuable insights. 
Trace Fissionable Uranium 

According to some authorities (15) the fissionable Uranium in one 
kilogram of granite (10 ppm) has a potential heat release equivalent to 



157 



32 
27«5 kg of coal. How then, are we to treat the energy requirement of 
a house built of granite? An average small (100 m ) built home will 
require about 50 m of stone. The specific gravity of granite is 2.5, 
hence the weight of stone is 125000 kg, which is equivalent to 3. k x 10 
kg of coal or about 100,000 GJ heat. Elbek and Petersen (ibid) have 
computed that the Gross Energy Requirement for a Danish house is about 
500 GJ. However, if both computations are based on Net Energy Requirement, 
that for the Danish house falls by a small amount (chiefly through 
the energy sequestered in plastic components and wood) while that of 
the granite built house falls to a value considerably lower than that 
of the Danish house. Which truly reflects the energy requirement for 
house building? 

^.10 Process energy requirement- 

Where the area of interest is strictly in comparing the energy 
sequestered in order to promote a given process, perhaps to compare it 
with another, it is only necessary to know the direct and indirect 
energy delivered to the process, and not the GER of the material inputs. 
Such a value is called the Process Energy Requirement PER. If a 
process was analysed right back to its primordial inputs, then, and then 
only, would PER ■ GER. 

Jf.ll Energy Requirement for Energy 

This value is extremely important in dealing with the exploitation, 
development or use of energy resources. It is the energy required to 
make a unit of energy available. It is therefore 



1/ 



energy in delivered fuel 



where PER fuel is all the energy used to explore, exploit, transport and 
process energy to the point of consumption, that is delivered fuel. It 
is a ratio, and will always be ^ 1. 

It is customary amongst many workers to refer to the above 

GER fuel " PER f 1 as the " ener S v cost of ener gy"- 

Note: NER fuel is taken as positive, though thermodynamic convention gives it 
a negative value. » | j :< 



158 



33 

4.12 Products as an energy source 

The foregoing uranium in granite example is distrubing for it 
shows that in certain circumstances different conventions can lead to 
vastly different answers. The problem partly arises through making 
calculations in terms of Gross Enthalpy, and not Free Energy. The energy 
actually available from the granite blocks of a house is considerably 
less than the potential heat that could be generated by the fission of 
the U atoms in the granite. The granite has to be mined and crushed. 
The uranium, which is present in minute quantities, has to be leached 
out, and selectively concentrated, and finally subjected to enrichment. 
The Free Energy for carrying out these processes has never, so far as 
we are aware, been calculated, but will certainly be very large. If 

this necessary Free Energy were substracted from the Free Energy in the 

235 
dispersed atoms of U, we should then have a better measure of the 

true energy sequestered by making a house in granite rather than in 

the more conventional materials of present day house technology. In 

subsequent correspondence the workshop participants considered this 

problem, and whether it called for a modified definition of GER. It was 

felt by most that it was sufficient to retain the definition as in 4.5, 

but necessary to be aware of the implications of the energy content of 

products and indicate the awareness in any published work. 

4.13 Summary of recommendations on terms of account 

GER Gross Energy Requirement . This may be expressed in terms of Enthalpy 
(GER) or Free Energy GFER. 

NER Net Energy Requirement . This is the GER less the gross energy 
(G or H as the case may be) of combustion of the product. 

PER Process Energy Requirement . The net energy requirements of all the 
inputs and stages of the process, summed over inputs and outputs. 

GER fuel Gross Energy Requirement for a fuel is the GER of total amount 
of that fuel input, whether consumed, rejected as waste, transformed or 



159 



3* 

stored in waste products that is required to furnish one unit of delivered 
energy. 

NER fuel . Net Energy Requirement of fuel is the GER fuel less a 
credit for any unconsumed energy in the waste products of combustion or 
fission. 

ERE Energy Requirement for Fuel . This value is the ratio of energy 
sequestered to deliver a unit of energy divided by the unit of energy. 
It is ^1. 

k.lk Units of account 

The workshop recommended that energy normally be expressed in 
terms of joules of power of ten thereof, and that where, for purposes 
of popularisation, it was thought necessary to use units like Tons of oil 
equivalent or Tons of coal equivalent, a footnote gives the value chosen 
in joules, since there is no absolute value. The workshop noted that 
it is customary to refer to a Ton of Coal Equivalent as 7 • Gcal per tonne 
or 29.28 GJ/tonne and a Ton of oil equivalent as 10 • Gcal/tonne or 
41.83 GJ/tonne. 

The workshop recommended that no results be expressed in British 
thermal units (Btu), since this unit was largely obsolete, and no longer 
fitted into any of the widely used units of measurement. 

4.15 Some symbols and descriptions in use at the time of the workshop, 
not now recommended. 

Energy cost Usually used to express the concept of Gross Energy 
Requirement, or something similar. 

Energy Equivalent of Network Input (ENI) This has been the GER or on 
occasion the "Apparent" Energy Requirement for an input to a process. 

Energy cost of energy This phrase usually connotates the value. 

Energy subsidy Usually akin to the Gross Energy Requirement. 



160 




Fig, 4.1 SYSTEM BOUNDARIES TO ESTIMATE G.E.R. 



161 

36 
SYSTEM BOUNDARY 



SOURCE 



/ 




/ 



COAL- 



FOSSIL FUELED 
POWER STATION 



I 

| COAL ■ 

. SUPPLY 

\ SYSTEM 

\ 

\ 
\ 
\ 



\ ^ 1000 Kwh 
\ * ELECTRICITY 




// OTHER 
— 7*— \ FOSSIL 
/ / FUELS 



/ 



S 



Fig. k.2 



SOURCE 



ENRICHED 
FUEL — 
X% 235u 



NUCLEAR 
( THERMAL) 

POWER STATION 



235ij (MfNING, TRANSPORT 

SUPPLY ENRICHMENT, DEUVERY) 
SYSTEM 



1000 Kwh 
ELECTRICITY 




Fig. U.2 

286 



162 



37 




BOTTLE 



Fig. k U 
MAKING A POLYTHENE BOTTLE FROM NAPTHA 



163 



38 
SECTION 5 



WASTE 



It is a common place activity in process engineering to assess energy 
usage in an industrial process, seeking opportunities to reduce such waste. 
The methodology of such analysis need not be described here, since it is 
embodied in chemical engineering texts. However one purpose that energy 
analysis hopes to serve is a better understanding of the relation between 
energy and economics and energy and production. Implicit in much economic 
thinking about the future is a belief that better technology will create ever 
more favourable relationships between production and resources. Thermodynamic 
analysis shows there is a limit to such improvement. Energy Analysis can 
play an important role by assessing processes and transformations in terms 
of Free Energy. 

It is a fundamental concept of thermodynamics that the Free Energy 
change to make product Y by the transformation of a number of inputs is related 
only to the initial and final states of the inputs and products. Where this 
is calculated as the change in Free Energy, ^G, of the materials appearing in 
the products this quantity is referred to as the ideal Free Energy change, and 
represents the minimum possible, irrespective of technology. Moreover 
such ideal Free Energy applies only to a process proceeding at an infinitely 
slow rate and may correspond to no known technology. All real processes 
must proceed at a finite rate, for which they must pay a AG penalty. 

It is useful to compute the ideal Free Energy change of a transformation, 
for it sets an ultimate limit not implicit in economic theory. This may 
then be compared with the Actual Free Energy change that was necessary to 
effect the transformation. In this way one has an excellent measure of both 
waste, and the margin for improvement. The now classic example of this 
was the automobile study by Berry and Fels (ibid). 

The workshop recommended that a Waste factor be defined as: 



164 



39 

Actual AG required to effect transformation - Ideal AG to effect transformation 
ActualAG required to effect transformation 

The ideal value is obtained by thermodynamic calculations upon the 

components of the transformation. Though not always easy to do, such a 

procedure is well understood and is backed by a plethora of text books and 

reference data. The acual G is obtained by computing the G of combustion 

(at standard state) of the fuels consumed in driving the transformation. 

This value will be extremely sensitive to the system boundary chosen. This 

is dealt with in Section 6. 

Waste factor as defined above can exceed unity only when AG. . .. ^-0 

J J ideal 

The closer the value approaches zero the less wasteful and more ideal 
is the process. 

Waste itself, is AG actual - AG ideal. 

Table 5-1 gives some waste factors from recent studies. 

TABLE 5-1 



Free Energy use, actual and ideal. 
(Sources: (l) Gyftopolous et al, Thermo-Electron, Waltham 
Mass, 197^.) 


(2) 


Berry and Fels - ibid. 




Product 




G actual for industry 
in US in 1968 

MJ joulesAg 


ideal 


Waste 
Factor 


Coking of coal (2) 








1.13 


Iron (1) 




25 


6 


.76 


Gasoline (1) 




k.2 


.k 


• 9 


Paper (1) 




38 


.2 


1.005 


Aluminium (l) 




190 


25 


• 87 


Cement (1) 




7.8 


.8 


.9 


Steel from Fe (2) 








1.19 


Bauxite - A1 2 (2) 








1.0 


Zinc smelting (2) 






1.02 



165 



SECTION 6 

SYSTEM BOUNDARY, DATA AND MAN- POWER 
% 

6.1 A system boundary Energy Analysis is a form of systems analysis, 
and as such involves a specification as to the boundary of the system 
under study. Fig. *+.l was an example of several possible boundaries. 
To probe the problem more deeply, let us suppose our task is to assess 
the Gross or Net Energy Requirement for a glass bottle. Fig. 6.1 
depicts a number of possible system boundaries we could choose. 
The inner (-----) would take into account only the direct 
energy used in making bottles, which would be very far from the real 
energy requirement for making a glass bottle. The outer ( -.-.-.-) 
requires an impractically large amount of information before either the 
GEE or NER may be evaluated. The workshop recommended the following 
scheme : 

Fig. 6.1 can be re-drawn so as to think of the process at 
decreasing levels of energy ijoput CFd-gi^.Z) . 



Level 1. Computing GEE or NER from this level alone is not 

sufficiently informative. In many cases direct and 
transport energy at level 1 may not even be 50$ of 
the total GER for the process. 

Level 2. Here the energy required to make the material inputs 

available is included - that is the GER or NER of those 
inputs. In a great many cases these two levels will 
comprise 90# of the total GER or NER. 

Level 3» Here we include the energy required to make the capital 

equipment for the process. Input-output analysis suggests 
that this is rarely more than 10# of the total GER or NER. 



166 



Level k. This takes into account the energy to make the machines that 
make the machines. For most systems this energy requirement 
will be below the noise level, and is almost certainly 
within the range of uncertainty of the final GER, NER or PER. 

Though the workshop considered the above scheme and Fig. 6.2 as 
a useful guide it emphasised that how far one took the analysis depended 
on the question asked. For example if the process under study was 
strip mining of coal, and one wished to include environmental 
considerations, some inputs at levels 3 or h might be quite large; for 
example, the energy requirement for land restoration. In dealing with 
the energy requirement for capital it was often not sufficient to 
consider only steady state system. Consider the following energy 
requirement for energy example in 6.1.1. 



167 



6.1.1 Energy requirements of energy - a paradox? 

Fig. 6.3 depicts a highly-simplified national energy system: 



stock 
system 



energy 
system 



E L 



acquire 
convert 
deliver 
etc 



/ 



production-consumption 
or user system 



£> E d 



J cd 



FIG. 6.3 



E is an energy stock, or fuel in the ground. As soon as it is 
disturbed by man it enters the Energy System where it is converted to 
useful or delivered energy (E ) and losses (E..). Part of E is diverted 
to search for, obtain, and make available the energy used in the 
production consumption system. Part of that diverted energy is required 



being furnished to tne consumer no further capital equipment is necessary, 
and the capital equipment is essentially outside the system boundary. 
In other words the system boundary is time dependent. In a money 
accounting procedure one would amortize the capital cost over a number 
of years, and add it to the running costs of the energy producing process. 
If, however, it was necessary to ascertain whether the money was available 
for building the energy producing system, it would first be necessary 
to secure a source of capital; similarly with energy. A certain 
amount of energy must be made available and then sequestered in capital 



168 



^3 

equipment before further energy can be made available. Fig. 6.k 
depicts the situation. Energy is invested in constructing (say) a 
nuclear power station. Once operating there is an energy requirement 
for operation, shown as a slowly declining cumulative EEE plot. 
However, once operating, the reactor has a rate of energy production 
resulting in a cumulative production (line c-d-). The net energy 
release into the consuming system is the difference between 
cumulative ERE and cumulative production. The viability of an 
energy producing system is not a simple matter of EEE being less than 
E (Fig. 6.3) but of time also. 

As an example of the power of energy analysis to reveal the 
picture in more detail, consider a country embarked on a programme 
of nuclear reactors. Suppose such reactors to have a 30 year life, 
and to have a capital energy requirement to build them equal to 10# 
of the total energy production in their life; that it takes six 
years to build a reactor, and one new reactor is started each year, 
Fig. 6.5 shows that for many years the country will incur a net energy 
deficit (11 years) and that 20 years will elapse before the cumulative 
energy production exceeds the cumulative energy investment. 

One must stress these figures are for illustrative purposes only. 
6.2 DATA 

6.2.1 The data required for Energy Analysis is very much a function 
of the level under consideration, as depicted in Fig. 6.2. 

Level 1 The data for Level 1 has to be obtained from actual process 
under study. Occasionally if the industry is extremely homogeneous 
and good economic data is available, an input/output table 
yields adequate average figures for the whole industry. Level 1 is 
the most highly disaggregated level. 



169 



UL 



Level 2 At this level the information may be more aggregated than 
in level 1. It may be enough to deal with the energy requirement 
of each significant material input, aggregating the lesser inputs and 
ascribing them an average value. Here again, the degree of dis- 
aggregation of data depends on whether one is looking at a whole 
industry or a particular sub-system of that industry. 

Level 3 Input/Output analysis suggests (l^t) that the energy requirements 
for capital are very much the same for all capital equipment in given class. 
Except where a sub-system is being examined in very fine detail the data from 
Input/Output tables will suffice. 

Level k When used, the data from Input/Output tables is sufficient. 



170 



*»5 

The workshop expressed a number of comments about data on 
energy requirements. 

It is dangerous to interchange consumption data with production data. 

* Many energy statistics are computed upon bases different from those 
selected by the workshop. Often these bases are unknown. 

* When quoting or reporting Energy Requirement figures, it is 
important that there be an explicit statement of whether they 
are Net or Gross, Enthalpy or Free Energy. 



6.2.2 Recommendations Regarding Data and Documentation 

We recommend that an international panel be set up to develop 
a set of conventions and standards for the definition', ^collection, 
validation and presentation of data on national energy production and 
use, the report of this committee to be sent to the appropriate national 
and international bodies for general adoption. 

We recommend that a working group be established to carry through 
a demonstration of the methods of documentation as stated in our 
specific recommendations following. Members of the present body are 
prepared to take active part in this effort. 

Specific recommendations on energy data sources for governmental 

statistics 

(1) National census data reports present energy in physical (S.I.) units. 

(2) National census data contain total production of each primary 
energy- supplying material in S.I. units of both mass and energy 
(gross heats of combustion for fossil fuels; total releasable 
heat from fissile nuclei for nuclear fuels; total developed 

electric energy for hydro-electric plants; total developed electricity, 
heat sold and heat used internally for geothermal stations). 



171 



he 

(3) National census data contain unit prices, price scales, or 
unit production costs for each energy- supplying material 
based on the same time-averaging method used for the physical 
data for energy analysts. 

CO Whenever a data base other than a national census is used, 

reference be given to the primary source of the data if it is 
at all possible, as well as to the immediate source from which 
the information has been drawn. 

(5) Each calculation of energy data be accompanied by an explicit 
statement of the assumptions and method of calculation. 

6.3 Man Power 

In Fig. 6.1 one common input to a manufacturing process was 
missing. That input was labour. In Energy Analysis should the energy 
equivalent of labour be included? If so, what is the energy equivalent 
of labour? 

Labour energy can be considered in two ways. One way, the metabol- 
izable calorie intake, calls for hugely different energy requirements 
according to the intensity of the agricultural system which furnishes 
the food. On the other hand where Energy Analysis is concerned with 
the use of energy sources and their depletion, the figure of real interest 
is how much in the way of energy sources was consumed to furnish the 
life-support system of the man that works on the process. This, one 
might say, is the Gross Energy Requirement of an individual worker. 
It was this second approach that the workshop supported. 

However, once this approach is accepted a further problem arises. 
Does one include in the energy for the life support of the worker, only 
food, or also his family, house, car etc? 

The problem is analysed by considering systems at various levels 
of development. The data for the following examples, though poor leads 
to firm conclusion. 



172 



^7 



6.3-1 Primitive agricultural system 

Fig. 6.6 depicts an agricultural unit (AU), totally isolated 
from any manufacturing economy - some Amazonian or New Guinea tribe 
living in isolation. 

All the product food, y, is consumed by the tribe. They purchase 
no inputs to the land, hence x. = 0. But they do consume a little 
firewood, x . . The energy analysis call for selection of some system 
boundary through which inputs enter and products leave. This is 
shown in Fig. 6.3 as the dotted line. 

6.3>1»1 Ignoring man-power 

x. 
The Gross Energy requirement for food = — = 0. 

That is, there is zero energy requirement to furnishing food 

to the tribe. Is this really true? The tribe make nothing 

for export outside their community, but they do, almost certainly 

consume firewood, which is an energy source, in their cooking 

and other activities. 

6.3'1«2 Including man-power 

Let x., the input to the households, be the firewood 
collected and used. One may get some feeling for the order of 
magnitude of this amount of energy by assuming a family size 
of five, each eating an average of 8.3 MJ/day (2000 k.cal/day) metab- 
olizable energy and a household wood consumption of 10 kg/day. 
If we assume that the only work done by the tribe is in food 
production, then all the energy of the house can be allocated 
to food production. 

Gross Energy Requirement becomes 
10 kg x 10 x 10 (JAg w °od) /K>,000 k.cal x *+lo"3 

that is 2.k joules energy resource per joule of metabolisable 
energy. This is not significantly better than much modern 



173 



agriculture (Pimentel, ibid) where the calculation 
takes everything into account other than the energy for life 
support system for labour. It is important to note that in the 
primitive system nothing crosses the system boundary of the tribe 
other than solar energy. 

6.3*2 Agriculture in an under-developed country 

Fig. 6-7 depicts an agricultural unit in an undeveloped 
country. Agriculture is at such an intensity that some food, a 
proportion f, is sold off the Unit, while the rest is consumed internally. 
The money obtained permits purchases of inputs both to the household 
x. and the land x. . 

6. 3*2.1 Without manpower 

Here we only include the energy equivalent of the inputs x. 
x. 
GER = — . In low intensity agriculture a typical input would 

be 1000 MJ/hectare.yr (15) and the production would be 

6000 MJ/hectare.yr metabolisable energy. Assume five people 

per household average each eating 2000 k.cal of metaboli able 

food energy per day, and assume that the whole family works on 

the land, which amounts to 5 hectares per family. 

Their metabolisable energy need is therefore 3*65 x 10 
k.cal/year, of which one fifth, should be allocated to each 
hectare, that is 305^ MJ. 

Hence (1 - f)» 6000 = 305^ , f = .**9 

GEE = 1000 / (.^9 • 6000) = .33 joules input energy per 

metabolisable joule of food 
energy. 

6.3.2.2 With man-power 

The hypothetical unit described above sold approximately 
3000 MJ metabolisable energy per hectare, having a present day 



174 



ho 



money value of around #250. Supposing all this money was 
spent on input having an average industrial country energy 
intensity of approximately 150 MJ/#, then the inputs to the 
household would be around 37,000 MJ. 

Supposing part of this money was spent on fertilizers 
equal to the 1000 MJ/hectare, that is 5000 MJ in all. This 
would cost about 70, leaving #180 for purchases of inputs to 
the household. The average energy intensity of purchases 
could be reasonably assessed on current values at 150 MJ/#, 
and therefore amounted to 27,000 MJ. 

If all the man-power, and therefore all the life support 
system for the labour is attributed to the food production, 
the Gross Energy Requirement becomes 

(1000 + 27, 000/5)/. *+9 x 6000 = 2.17 joules input/joule of 

metabolizable energy. 

3.3 Low intensity developed agriculture - hill farm 

Fig. 6.8 depicts a U.K. hill farm system with typical inputs (16). 

6.3*3*1 Without man-power 

GER = JfOO GJ/10,000 kg protein = 400 MJAg protein. 

6.3*3*2 Including man-power 

A typical small household purchases in direct and indirect 
energy of one sort and another about 80 GJ/year. We have two 
choices. In the hill farm as small as the one described, one 
man will work it. Do we ascribe all the energy inputs to house- 
hold to that one worker, or merely that proportion which supports 
him? If we included the total, arguing that his family was 
part of the life support system for the man, then 

GER = C+00 + 80) GJ/10,000 kg protein = k80 MJAg protein. 
Including man-power on_this basis has raised GER by 20#, somewhat 



175 



5C 



larger than any likely errors in a well-founded analysis. 

If, however, we argue that in this family of five, only 
one fifth should be attributed to the worker, the inclusion of 
man-power then falls into the noise level of available data on 
other inputs. 

6.3*^ Factory process 

Fig. 6.9 depicts an ammonia factory process. The input 
energy is taken from published values (17) 



6.3«^«1 Without man-power 

GER = ifO,000 GJ/ 10 6 kg = kO MJAg 



6.3«^«2 Including man-power 

Let us assume a factory with 200 men over the three shifts 
and including maintenance and management. Their gross life 
support system is 200 * 80 GJ per year or ^3.8 GJ/day. If all 
this is ascribed to the process 

GER = C+0,000 + ^3-8 GJ) / 10 6 MJAg 
man-power becomes unimportant whether included or not. Even if 
it be argued that the industrial worker uses more inputs to his 
home than the agricultural worker, even that will not appreciably 
alter the result. 

6.3.5 Workshop recommendation on man-power 

WHERE THE ANALYSIS REFERS TO DEVELOPED OR INDUSTRIALISED ECONOMIES 
IT IS NOT NECESSARY TO CONSIDER THE ENERGY FOR LIFE-SUPPORT OF MAN-POWER. 
WHERE THE ANALYSIS CONSIDERS LOW INTENSITY AGRICULTURE MAN-POWER CONSIDER- 
ATIONS PLAY AN IMPORTANT ROLE IN THE CALCULATIONS. 

However, the problem of partition between household and labour or 
labour and its life support system was a matter the workshop did not 
resolve. A convention still has to be established. 



176 



51 



TRANSPORT 




Fig 6.1 BOTTLE PRODUCTION SYSTEM SHOWING SOME 
ELEMENTS OF THE TOTAL SYSTEM. 



177 



52 



LEVEL 1 


LEVEL 2 


LEVEL 3 


LEVEL U 


DIRECT E 
TO PROC 

> 


ENERGY 
ESS 


DIRECT ENERGY 
PROCESS 


DIRECT ENERGY 

1 


DIRECT ENERGY 










ALS 


RAW MATERIALS , 


'MACHINES TO 
MAKE 
MACHINERY 










MACHINES i 
(CAPITAL EQUIPMENT) 


* 




i 




i 










i 




ENERGY 
TRANSPO 


OF 
RT 


ENERC 
TRAN! 


Y OF 
5PORT 


ENERC 

TRANSF 


r of 

'ORT 


ENERGY OF 
TRANSPORT 


GER 
or 
NER 


GER 
or 
NER 







Fig. 6.2 



178 



53 




Fig. 6.4 ENERGY PRODUCTION AND USE PATTERN OF A 
NUCLEAR OR OTHER ENERGY CONVERTER. 



179 



54 











3 








co 






1 


UJ 








§g 






UJ 


CD CO 






3: 


z ce 






H 


Q 5 

CD 




XX. 

ecu) 




UJ **■ 








e> uj 

lo i 




o 




_J 






or 




ooc 




P CO 




«o 




O ^ 




1 

g 




< 5 • 






UJ >- Q 




£ 




or o -• 




in 




co 3 




,*, 




or co 




0_ 




UJ it 
_l t o 








O £ H 




L § 


_ 


3 




t— 




Z -»- 




i ^ 




or z> 




1 z 




^- < o- 

<UJ^ 

z o 


> 

UJ 






o or 


— "~ C\> 


\Vs 


or uj >_ 
u- o_ g 


II 

00 Q. 






3 or UJ 






p- o z 






E 1- UJ 








O ° 






HN 


2 LO 






K\ 


POWER 

HER F 

3 YEAR 








NET 

FURT 

AND 








\\^ xc> 










v\vb ^ 










\VVr ^ 
VoJ u_ 





H3M0d 13N 



180 



55 



\ 



/ 
/ Inputs Xi = 



/ 









\ 

\ 




AGRICULTURAL 
UNIT 




\ 






Food, y ^ 
















/ 








/ / 








/ / 




Labour 




/ / 
/ / 

/ S 
/ y 
1 — 






/ 










/ 
/ 
/ 


TRIBE 




Inputs Xj 






/ 


= Wood 






S 











PRIMITIVE, ISOLATED TRIBE 



Fig. 6.6 



/ 

inputs Xi 



\ 



\ 



AGRICULTURAL 
UNIT. 



-»- Food y 



(1-f)y 



/ 



HOUSEHOLD 



SIMPLE AGRICULTURE;, SOME EXPORT OF FOOD Fig. 6.7 



181 



56 



y 



/ 



/ 



400 GJ/yr 



\ 



\ 



s 



Farm inputs 




10,000 Kg protein / yr 



Labour 



/ 



/ 



HOUSEHOLD 



80 GJ/yr 



LOW INTENSITY DEVELOPED AGRICULTURE 



Fig. 6.8 



£0.000 GJ/day 

/ 
/ 

/ 
/ 
/ 
I 
\ 
\ 
\ 
\ 



/ Factory inputs 



Labour 
'n' units 



\ 



N 



AMMONIA 
PROCESS 



\ 



-\ *- 10° Kg /day 



s 



/ 



WORKER 
HOUSEHOLDS 



80 GJ/ 



household yr 



INTENSIVE INDUSTRIAL PROCESS 



Fig. 6.9 



182 



57 



SECTION 7 

THE PROBLEM OF PARTITIONING ENERGY REQUIREMENTS 

7«1 Partition refers to the problem of analysing a system which 
simultaneously produces several goods or services. Even when the 
system boundary has been well defined and all the inputs counted with 
appropriate energy there sometimes remains the problem of dividing the 
sum of all in-puts between the outputs. 

An obvious example occurs in calculating the energy cost of road 
transport. Here it is fairly easy to calculate the energy needed to 
construct, light, clean and maintain the road - but how is this energy 
input to be divided between all the road users. What proportion of the 
input should be set against private motorists, what to road haulage by 
lorries and what to bus operators? 

Other examples are: 
Oil Refinery. What criterion does one choose to partition the energy 
requirements for distillation between products and wastes? 

Chlorine. If Chlorine is made by electrolysis of sodium chloride, 
caustic soda is produced in equi-molar quantities. Chlorine may be 
the wanted product, but caustic soda also has a value. Does one 
allocate all the energy requirement to chlorine? If not, how does 
one decide on the partition? 

7.2 Possible Conventions 

There are many conventions available to resolve the problem, 
most of which lead to different answers. The workshop considered the 
following the most tenable. 

Convention 7.2.1 Assign all the energy requirements to the output 
of interest. Hence in an analysis of road 
freight transport it is assumed that the GER 



183 

58 
of the road is all counted as part of the 

energy requirement of road transport. 

Convention 7*2.2 Assign energy requirement in proportion to financial 
value or payments. Under this convention the 
proportion of energy charged to road freight 
transport would be in proportion to their con- 
tribution to total vehicle taxation. 

Convention 7*2.3 Assign energy requirements in proportion to some 

physical parameter characterising the system. Thus 
energy requirements may be divided so that all the 
outputs have the same GER per unit weight, or per 
unit volume. In the example of transport energy 
the division could be in proportion to the number of 
vehicle-miles, or vehicle ton-miles. 

Convention 7-2.^+ Assign energy requirements in proportion to the 

marginal energy savings which could be made if the 
goods or service were not provided. Thus if one could 
calculate how much smaller roads could be and how 
the frequency of - maintenance could be reduced if 
road freight transport were reduced then this should 
should be proportional to the energy requirement 
chargeable to road freight transport. 

7«3 Workshop Recommendations on Partition 

Convention 7- 3-1 will produce a biased result, over-weighting the GER 

of a good or service, but sometimes useful in revealing 
an upper bound of GER if by-products ceased to be useful. 

Convention 7«3-2 seems inappropriate in an energy analysis, since it 
brings in money. However, the workshop recognised 
that relative money values of different products 
do reflect society's current perception of the 



184 

59 
relative values of products and wastes. This 
convention might, for example be as useful a way 
as any for partitioning energy requirements when 
considering desulphurisation of oil or gas. The 
disadvantage of this convention was that any GER 
or NER value derived would change with time in a 
manner unrelated to technological improvements. 

Convention 7.3*3 w as the convention recommended by the workshop for 
use whenever possible. Once again it should be 
appreciated that the choice of physical value 
should be as far removed as possible for a value- 
judgement. Thus one would partition in a petro- 
chemical plant on the basis of the enthalpy of combustion 
of the products. 

Convention 7«3«^ This convention confuses energy analysis with 

policy analysis based upon energy analysis, and was 
not recommended by the workshop. 



185 



60 



SECTION 8 

PROCEDURE FOR ENERGY ANALYSIS THROUGH PROCESS ANALYSIS 

8.1 The workshop examined process energy analysis through the 
procedures of process analysis. It recommended the following procedure, 
terminology and flow-charting. 

8.2 PROCEDURE 

The procedure for energy analysis falls into five parts. 

1. Decide objective of analysis, and therefore whether 
information is wanted in terms of NER or GER expressed as 
Enthalpy or Free Energy. 

2. Choose a system boundary. This is facilitated by 
making a flow sheet depicting inputs and outputs. 

3- Identify all inputs. 

h. Assign energy requirement (NER or GER) to all inputs. 

5. Identify all outputs. 

6. Establish criteria for partition. 

It is rarely possible to select the correct system boundary at 
the first trial. It may require a trial calculation to reveal the 
relative importance of various inputs. A lay-out such as recommended 
in Fig. 6.2 will assist. It is important to note that the system 
can be treated like a 'black box' only provided the system boundary 
is clearly understood, clearly delineated, and all inputs and outputs 
are counted as they cross the system boundary. Fig. 8.1 depicts 
some arbitrary production system. 

The Gross Energy Requirement of the products is the sum of the 
Gross Energy Requirement of the inputs, partitioned in a rational way 
(usually by convention 7- 3*3) amongst the products. 



1-391 o - 76 - i: 



186 



61 



■3 Terminology 

8.3»1 Abbreviations 



Recommended 



Abbreviation dimension 

* Gross Energy Requirement (defined 

(defined sectinn 4.8.1) GER or GFER MJ/kg 

* Net Energy Requirement 

(defined section 4.9.2) NER MJAg 

* Process Energy Requirement 

(defined section PER MJ/kg 

* Energy Requirement for Energy 

(defined section 4.1.1) ERE MJ NER of fuel/ 

MJ delivered 

However not everything has an explicit weight, this particularly applies 

to services. 

Historically, many studies have been made and published in terms of 

the Apparent Energy Requirement which is GER - NER (fuel). The workshop did not 

recommend an abbreviation since it was felt that to publish Apparent Energy 

Requirement values would be confusing, and misleading, and should be discontinued. 

8.3.2 limensions of energy units 

The workshop recommended that whenever possible energy should be expressed 
in joules (J) or multiples of powers of ten thereof, while the unit of product 
should be the kilogram or tonne (metric ton). There should be a clear state- 
ment of whether the resultant figures are in terms of Enthalpy (H) or Free 
Energy (G), and suggested that this could be indicated by Subscripts of H and G. 

8.3*3 Definitions 

The following words have precise meaning in Energy Analysis. 
Enthalpy (H ) This is the Gross Enthalpy of combustion in air at one bar 
pressure and 273-15K (see section 4 and table 4.l). If the energy source is 
nuclear, the gross heat (after fission product cooling) of the fission at 
the same temperature and pressure (see section 4.7.1). It is often referred 
to as "Higher Heating Value" or Gross heat of combustion. 



187 



62 
Free Energy (G) The usual meaning in thermodynamics, and related to 
Enthalpy by G » H - TAS. 

Waste Factor No symbol proposed. Definition given in section 7. 

Partition When a process requiring energy produces two or more 
products, users or services, the energy used must be allocated among 
them. This allocation is referred to as partition (see section 7). 

Direct Energy Enthalpy of fuels used directly in the process under 
study plus the electrical energy delivered to the process. 

Energy Intensity The energy requirement in terms of energy per unit 
money value of product. For example MJ/swedish kronor GJ/£ and so on. 

Where no qualification is given, energy will be in terms of GER 
and the money value in terms of the final production price. However, it 
is possible, by qualifying the number, to express energy intensity in 
terms of GER, NER, PER or even Apparent Energy Requirement, and to express 
the money value in terms of value added in the process. 

Delivered Energy The energy (expressed as H or G) delivered by an 
energy producing system to a consuming system. ERE will represent 
the energy requirement to make this delivered energy available. 

S.k Flow Charting 



Where a process is under study, and the author wants to reveal 
clearly all the components that add up to PER, a flow sheet is valuable. 
Conventional chemical engineering charting does not suffice for the intric- 
acies of Energy Analysis. The workshop therefore recommended a procedure 
with appropriate symbols. The purpose of both is to enable one set of 
calculations to be readily compared with another. The chart should state 
clearly whether numbers are in terms of enthalpy or Free Energy. The 
flow chart evaluates the Process Energy Requirement PER, not GER of 
product. All values are normalized to one unit of output. 



188 



63 
8.4.1 Flow chart symbols 

Flow chart symbols are shown in Fig. 8.2 and are largely as orig- 
inally proposed by Berry and Fels (ibid). 

* Name of process stage in rectangle. It is convenient to write in 
here the name of each stage in the process (e.g. milling, distil- 
lation, or the proprietary name of the process such as 'Bayer'). 

* GER of fuels: place in the triangle the number of joules GER 
fuels per unit of (not electricity) output. 

* Transport: place in the 'cart' the number of joules GER 
allocated to transport in this section of the process. 

* Capital: place in the 'diamond' the number of joules GER for 
capital, specifying in the text amortised over the life of the plant 
and per unit product. 

* Material: place in the circles the amount of kg or tonnes and 
name of material. 

* Electricity. The fuel symbol is modified as shown in Fig. 8.2. 

At the top is placed the kwh of electricity per unit of product. In the 
triangle place the total GER of the fuel needed to furnish that electricity. 

.* Enthalpy content of inputs. These are expressed as a number of joules 
per unit product inerted into an upper semi-circle. Normally used 
for combustible inputs. 

* Enthalpy content output. This number is placed in a lower semi-circle. 

It should be noted that the total NER of the process is then obtained 
by summing all values from the top down to the final product. This value 
is not GER unless the analysis goes back to resources in the 'ground' because 
the flow chart does not provide for the GER or NER of inputs other than 
direct energy. The flow chart therefore is primarily to demonstrate the 
i^^ess Energy Requirement of the process, PER. 

Fig. 8.3 gives an example for an aluminium process. 



189 



&* 



INPUTS Ex { GER, 



\ 



\ 



H \> 



/ 



OUTPUTS Exj GERj 



Fig. 8.1 COUNT ONLY FLOWS AS THEY PASS THROUGH THE 
SYSTEM BOUNDARY. 



190 



6s 



NAME OF 
PROCESS STAGE 



AH INPUT 




G£.R. OF 
FUELS 



TRANSPORT 



CTD 



CAPITAL 





QUANTITY 
"ELECTRICITY 



TOTAL G.E.R. 
FUELS 




Fig.8.2 SYMBOLS OF ENERGY ANALYSIS 






191 




Fig. 8.3 

ALUMINIUM PROCESS 
FLOW CHART FOR 
ENERGY ANALYSIS 



10 




ALUMINIUM 
1 TONNE 



192 



67 
SECTION 9 



INPUT-OUTPUT ANALYSIS 



9-1 Description 

Process or Vertical Analysis traces back all the inputs to a 
particular production system. The effort can result in very accurate 
data, not necessarily representative of a whole industry. It is also 
extremely time consuming. A swifter but highly aggregated answer may 
be obtained using the standard economic input-output tables which have 
been prepared for virtually all developed economies, usually through 
government statistical services. That of the U.S.A. is the most 
highly disaggregated, comprising some 368 sectors, while the U.K. is 
divided into 90 sectors. Thus if the objective of the energy analysis 
is to establish the Gross Energy Requirement for a kilogram of pineapples 
input/output tables are of no help whatsoever. But if the objective 
is to find out the average GER of the whole agricultural system, the 
I/O tables will provide the information swiftly usually to a high 
degree of accuracy. 

The published I/O tables are couched in money terms. 

9.2 Input-Output Theory 

The following borrows heavily from the publication of Herendeen 
and Bullard (1*0. 



193 



Money Based Input-Output Theory 

The basic data for I/O analysis are money sales per year measured 

in producers' prices. All quantities will be in these units unless 

otherwise stated: 

n 

Let X. =« total output (i.e. sales) of sector i. X. = 2_ X. . + Y. , 
1 X 3=1 1J X 

where X. . is an intermediate output (inter-industry sale) from i to j , 

and Y. is the final demand (sales) of i. The sum runs over all indust- 

n 
rial and commercial sectors. / X.. is the sum of all purchases by 

.s the value added (wages, depreciation, profits, 



7" X.. i £ 



Input-output theory now assumes that the inputs of a sector are 

X. . 

linearly related to the output. Thus -=-*■ = A. . (A. . is a dimensionless 

X. ij xj 

constant), and 



X. = Y A. . X. + Y. , 

i • -, i.l .1 i 



or, in matrix form 



This may be invertec 



A X + Y . 



X * (I - A)" 1 Y 



Three "tables" (matrices) are commonly used in input-output work: 



1. X is the transactions table (of money flows) for a year. 

In addition to the X. ., the final demands and total outputs 
are also tabulated. All are empirical quantities. 



2. A is the table of direct coefficients. The entries are 
computed from A. . * X. ./X . A. . is the money transaction 
from i to j required for j to produce a unit money's worth 
of total output. 



194 



3. (I - A)~ is the table of total (direct plus indirect) requirements. 

a a ^^^^— f 

(I - A) . . is the total money output of i required for the economy 
to deliver unit money's worth of j to final demand*. 
Assuming linearity, and, in addition, time invariance of the A. ., 
one may use the total coefficients to predict the total outputs 
required to supply an arbitrary final demand vector. This is 
the whole intent of use of input-output. 

As an example, consider the transaction table for a three 
sector economy: 



Crude Oil 


Crude Oil 


Refined Pet 


Cars 


TFD 


TO 





10 








10 


Refined pet 


5 


5 


5 


25 


ko 


Cars 


o 








20 


20 



where TFD * total final demand, TO = total output'. 



and 



(I - A)" 






Xfh 





V2 


1/8 


iA 








. 


7/6 


1/3 


1/12 


2/3 


V3 


V3 








1 



*(I - A)~ is defined in terms of final demand, while A is 

as = 

defined in terms of total output. However, it is rigorously true that 
■1 



(I - A)" 



-V 



One may verify that 
10 

ko 

20 



(I - A)" 



Note that 01 of TFD for a car requires #0.80 output from crude, #0.33 
from refined, and 01 from cars. 



195 



70 



9»3 Energy Based Input-Output Table 

An energy balance applied to individual sectors of the input-output 
tables yields equations similar to those in economic input-output analysis. 
Energy based input-output tables were first developed by Herendeen (ibid) 
at Oak Ridge National Laboratory, and completed at the Centre for Advanced 
Computation at University of Illinois at Urbana and refined through 
two further years work. The methodology described here is that of 
Herendeen's latest version (1*0, and refers to the 368 sector U.S. 
input-output tables. His values are in U.S. dollars and British Thermal 
Units (Btu). 

Fig. 9-1 depicts the energy balance of an economic sector. 

It is assumed that the energy required (that is the Gross Energy 
Requirement to that point) in the inputs to sector j, plus the energy 
consumed in that sector, is passed on as part of j's output. This is 
simply in accord with the laws of energy conservation. There is one 
equation for each of N sectors. By assuming there is a unique value, £ ., 
of the energy intensity (see section 8.3«3) associated with each sector's 
output, one can solve the system of linear equations for the vector £. 

Appendix 3 carries a full analysis by Herendeen of his methodology. 



9. 3-1 Units of input-output energy tables 

The energy values of the tables are Gross Energy Requirement, 
and theoretically embodies infinite regression. That is to say the 
input-output table almost always takes the energy analysis back to be 
fourth level and beyond (see section 6.1). However there are inaccuracies, 
which stem from: 

* Aggregation of input-output sector, which may embrace fairly 
diverse activities. 

• The problems of converting- money units to physical units - 
there are widely different purchase prices. 



196 



-I 



* Referring all data to the same time period. 

National input-output tables in energy effectively state 
total GER of country m J-. Y 



S.k Energy Requirement for Energy 

The input-output Energy Table is an ideal vehicle for 
establishing the Energy Requirement for Energy (ERE). Table (.1 
lists the U.S. input-output values against those computed for 
the U.K. by process analysis (15). 



TABLE 9.1 



Energy Requirement for Energy, 
Energy units per unit of energy made available. 


Coc\l 


USA from input- output 
table, 1963 (1*0 


UK, process analysis 
(15), 1963 


1 .OlJf 


1.0^7 


Refined Petroleu 


n 1.19 


1.23 


Electricity 


2.89 


3.5^ 


Natural* Gas 


1.16 


- 



EiX; 



197 



72 



^ E| Xy ♦ Ej earth = E } X 
i=1 

£ X * E earth = E X 



E = E earth (X-X)" 1 



Ej earth 



E]Xj 



Fig. 9 1 



198 



73 



IMPORTS 



EXPORTS 



ACTUAL (9-4% 
PENALTY 



EMBODIED 



U-7%) 





U S ECONOMY 
1963 








*" 






, 







ACTUAL (3 -3%) 
PENALTY 



EMBODIED 



5-3%) 



DOMESTIC (85-9%) 



U.S. ENERGY BALANCE 1963 



Fig. 9.2 






199 

SECTION 10 

MARGINAL AND AVERAGE ENERGY REQUIREMENT 

10.1 With its discussion of marginal and average energy costs the 
workshop began to share its vocabulary with economics. 

The words "marginal" and "average" energy requirement are 
intended to be used in an analogous sense to the words marginal and 
average cost, as expressed by an economist. 

If one considers a large manufacturing unit, or a whole 
industry, the average cost could be calculated by considering the 
total cost of the products, and dividing that by the total number of 
products, both expressed over the same time period. The marginal 
cost is the extra cost of making the last unit produced or the cash 
saved in making one less. Clearly the same concept may be applied 
to the energy requirement of a product. 



200 



75 



10.2 MARGINAL ENERGY REQUIREMENT 

Consider a production system with inputs expressed as energy- 
equivalents. 

E = direct energy requirements 

F = energy of feed stock ( H of combustion) 

M = energy required to make available the material inputs to 
production 

G = energy required to produce the capital plant, amortised 
over the expected production of the plant 

T = energy requirements for transportation 
Thus GER = E + M + C + T + F 
and 

PER = E+M+C+T 
Let q be the number of units of output per unit time. 

The usual energy analysis procedures produce average values of 
these two quantities: 

GER/q = (E + M 4 fP. + T + ?)/q 

PER/q = (E + M C + T)/q 

Energy analysis (expressed in economic terms) generally calculates 
the inverse of the average physical productivity of energy or the average 
energy requirement. This need not be so. It would be perfectly 
feasible to examine a production system to ascertain the marginal energy 
requirement, that is to say, the additional energy to make one more 
unit, or the energy saved by making one less unit. The marginal and 
average energy requirement will rarely be equal. Fig. 10.1 depicts 
the change in cost of a product with production rate (q) and Fig. 10.2 
expresses the same concept in terms of energy requirement. 

MTC = marginal total money cost MER = marginal gross energy req 

ATC = average total money cost AVER = GER = average gross energy req 



201 



76 



Under pure competition, a firm will produce an amount a at the 
minimum of its average total money cost, where MTC = ATC. For this 
production level, MER is not necessarily equal to AVER. There are four 
different marginal energy requirements: 



GER | (OPER 

~5T 



(* 



short run marginal 
M,C,T,F V* ' /M,C,T ener ^ requirement 



d GER _ d PER long run marginal 

dq dq energy requirement 



The use of marginal and average energy requirement should be 
explicitly stated in any reported work. An especial caution is sounded 
about using values of average energy requirement in a marginal sense in 
any discussions. 

10.3 OBTAINING ACTUAL VALUES 

The input-output tables provide average and not marginal energy 
requirements though it is possible that when a series have been produced 
for different years the change in average cost may permit calculations of 
marginal energy requirement for whole industries. 

The marginal energy requirement can only be obtained at the present 
time by a study of an actual process, firm or industry in some depth. 
Such studies have only recently been commenced. 



68-391 O - 76 - 1' 



202 



7? 



MC=ATC : ^ 3 




1. 

production rate 
Fig. 10.1 




production rate 
Fig. 10:2 



203 



79 



SECTION 12 



INTERFACE BETWEEN ENERGY AND ECONOMICS 



In its final session on 31st August, 197^ the workshop considered 
what the next step should be, in particular how energy analysis could aid the 
understanding of our economic system. It was unanimously decided that a 
further and larger workshop or conference should be convened to consider the 
interface between energy analysis and economics. 

This conference would seek to bring economists and energy analysts 
and others together. Several areas of interest were discussed, in particular 
the role of energy analysis in economic allocation, energy use as a function 
of production haste, and energy analysis as a signal for impending change. 
Several other likely areas are embodied in Section 2. 

It was hoped that this conference would be convened and sponsored 
through IFIAS. A four man steering committee was nominated whose role is 
to 

* contact interested persons, 

* seek a venue and find sponsorship for the cost of the conference, 

* select a programme for discussion, 

* invite papers. 

Interested persons are encouraged to contact the committee representative 
in their area. (Addresses and telephone numbers are given in the list of 
participants on page 3) • 



The Americas: Dr Frank Alessio, Electrical Power Research Association, 
Palo Alto. 

Japan and Asia: Dr Takeo Hoshi, Tokyo. 

Continental Europe: Professor Bent Elbek, Niels Bohr Institute, Copenhagen. 

British Isles: Dr Malcolm Slesser, University of Strathclyde, Glasgow. 
(Chairman) 



204 



80 

APPENDIX 1 

Notes' Some comments on Energy sources. 

1 . By specifying standard reference states, and by use of the 
term "natural", we remove exotics like hydrogen from the 
class of resources. 

2. Reference states for substances could include 

a) temperature 

b) pressure 

c) concentration 

d) chemical formula (e.g. silicon as SiO_) 
l nuclear species 

f) magnetization 

g) etc., etc. 

3. Certain substances like Cu, Fe, S occur in small 
amounts and do qualify as resources under the definition, 
but "practically" speaking are negligible. 

4. The energy resources of the earth appear finite. 

5. Reference levels for fluxes could include* 

(a) Sea level for water-power 

(b) Some radiation temperature (3°K for free space?) 

6. Strictly speaking, there is little fundamental difference 
between energy resources and energy flux sources. For 
example, geothermal energy will be worked as a resource, 
i.e. it will be "mined". 



205 



APPENDIX 2 



31 



The thermodynamics of available work 

The section is adapted from the Energy Conservation Study 
carried out for the NSF Office of Energy R&D Policy by the 
American Physical Society, Summer 1974. 

Consider a system characterized by Energy E, entropy S, 
and volume V (see Fig.A). Its energy may include, in addition 
to its internal energy, also gravitational potential energy, 
kinetic energy of bulk motion, etc. 



Heat to 
atmosphere 




Figure A. Schematic diagram of system interacting with the atmosphere and 
transferring work to other systems. The available work of the system is the 
maximum useful work it can transfer to other systems as it comes to thermodynamic 
equilibrium with the atmosphere. 



206 



82 



(Temperatures, pressure, and other intensive variables may vary from one part 
of the system to another. ) Now let the system proceed via chemical reactions 
and/or other changes until it comes into thermodynamic equilibrium with the 
atmosphere. (By "atmosphere" we mean an appropriate large reservoir comprising 
the environment of the system; usually it will in fact be the earth's atmosphere.) 
The atmosphere has temperatures T , pressure P , and a specified composition. 
The system, as it changes, may exchange heat and work, but not matter, with the 
atmosphere, and it may also do work on other systems. The latter is called 
"useful work". Upon reaching equilibrium with the atmosphere, the system has 
energy E-, entropy S f , and volume V_ . The maximum useful work that can be 
transferred by the system is called its available work. The available work is 

determined by the condition that the total entropy of the system plus the 

do xxx 

environment/ not change (i.e., that all processes are carried out reversibly ). 

In the subsection below we show that for the conditions stated, the maximum 
useful work is 

A = (E-E f ) + P Q (V-V f ) - T Q (S-S f ) . 

[available work without diffusion] 
The second and third terms on the right represent energy "bootlegged" from the 
atmosphere. If the system's volume decreases (V-V f > 0), the atmosphere does 
work on the system, which can be passed on as useful work (the second term is 
then positive). If the system's entropy increases (S-S f < 0), and the increase 



We generalize to matter exchange (diffusion) later. 

As noted earlier, this quantity is frequently called "availability". 

Reversible change may require that one or more Carnot engines be interposed 
between the system and the atmosphere; see Fig. 2.U. 



207 



83 

is accomplished reversibly, the atmosphere transfers heat to the system, which 

can be passed on as useful work (the third term is then positive). 

The available work is closely related to the change of the Gibbs function, 

or Gibbs free energy (G = E + PV - TS), but B differs from -tfj in that the 

definition of B includes the temperature and pressure of the atmosphere only, not 

the system. This means that B can be defined for a nonequilibrium state. In 

one important special case, that in which the "system" consists of a quantity 

of fuel at atmospheric pressure and room temperature, the available work and the 

Gibbs free energy change are the same (except for the small additional diffusion 

term), 
atmosphere?); there is the problem of the logarithmic divergence; and there 

is the practical impossibility of harnessing much, if any, of the diffusion 

contribution to available work. For these reasons, we recommend that the 

restricted definition of available work (Equation 2.10) be used and that the 

efficiencies of all power-producing devices be expressed in the form 

e = V out/ B - 

The Combustion of Methane: An Idealized Calculation 

The combustion of methane in air can be represented by 

CH U + 20 2 + 8N 2 -> C0 2 + 2H 2 + 8N 2 . 

enthalpy 
The net/ of combustion is 192 kcal per mole of CH. . 

The adiabatic flame temperature is 22U0 K (see, for example, 

Van Wylen, 1959). It is necessary to calculate 

the entropy contribution to available work exactly and to calculate approximately 

in an ideal-gas approximation - the flame temperature (confirming the given 

value) and the loss of available work in the combustion process. This will 

show the relative magnitude of the four contributions to A in a 

practical case. Because of the ideal-gas approximation, it will provide 

insight that is perhaps not obtained in more exact treatments. And it will 

reveal the substantial loss of available work associated with the irreversibility 

of combustion. 



208 



84 

Adiabatic combustion carries the initial fuel-air mixture in state 1 
to the products in state c (see Figure A ). Subsequent processes (not 
considered here) carry the products to a final state f in equilibrium with the 
atmosphere. The energy that goes into heating and expanding the products in 
the flame is 

-AE = -AH + p ( V f~V ' 
where -AH is the heat of combustion. In this particular example, V =V, 
because these are 11 moles of gas in both state 1 and state f. Therefore, 
in an ideal-gas approximation, the temperature of state c is determined by 
the energy equation 

-AH = nC^(T c -T Q ) , 

where n is the number of moles (ll) and C p is an approximate average specific 
heat at constant volume for the products. It is helpful to express C p as a 
multiple of the gas constant R and to express -AH as a multiple of the thermal 



The dimensionless specific heat a can be estimated to be about 4.5 (it would be 
3.5 for a diatomic gas near room temperature). The dimensionless enthalpy 
is h = 324 for T Q = 298 K (RT Q = 0.592 kcal/mole). Then Equation 2-37 takes 

the form 

T 



'o 

solving for the flame temperature, we get 



In this example. 



h = na( ^ - l) ; 
o 

, we get 

T - T (l + -M . 

c o\ na/ 



h = 324, n = 11, a ^ 4.5, ^ = 6.55 . 

These numbers lead to T = (298) (7- 55) = 2,250 K, in close agreement with the 
actual value (2,240 K). (it can be remarked, incidentally, that the flame 



209 



25 



temperature for other hydrocarbon fuels will not differ greatly from this 
value because h and n scale in approximate proportion to the molecular weight 
of the fuel while a is approximately the same for all fuels.) 

Now we look at available work. In adiabatic combustion, the entropy of 
the atmosphere does not change. We can therefore write (from Eq. 2.20), 



where AS is the entropy change of the atmosphere plus the system and AS is 
the entropy change of the system. This can be written 

AS = AS, (from change of composition) 

+ AS ? (from isobaric heating of products) . 

The first of these two contributions to entropy change can be further subdivided: 
AS, = AS, (from summing individual component entropies) 
+AS, h (from mixing of components) . 

The second contribution, in an ideal-gas approximation, can be written 

AS 2 = nc; in(T c /T o ) . 



In this example, we find 



T o AS la = "°' 36 kcal / mole " -0.6lRT , 



T Q AS 2 = 59- h kcal/mole = +100RT Q . 



Total, T AS ^ 99RT . 
OS o 



The available work left in the flame is 



A = A - T AS 

c o a 



210 



36 




@ 



Combustion 



Heat 




State 1 



State c 



State f 



Figure Schematic view of combustion. Process oL is adiabatic combustion. 
The energy released goes into heating and expanding the products. After sub- 
sequent processes G> , the products reach volume V- at ambient conditions of 
temperature and pressure. 



211 



8? 



Consistent with the present level of approximation, we can set a 2= -AH = hRT . 
Then Equation 2.U6 takes the simple approximate form 

For the specific numbers of this example (h/na = 6.55), 

A 2: O.69B ; 
c 

this means that 31 percent of the available work has been lost in the combustion 
process. 

In a similar but probably more exact sample calculation, Keenan et al (1973) 
treat a hypothetical fuel CH with a heat of combustion of 20,000 BTU/lb and 
find |m( to be 29 percent of the heat of combustion or 27 percent of the initial 
available work. Applying the simple approximations of this example to other 
hydrocarbon fuels, we find values of |m| ranging from 29 to 31 percent of the 
heats of combustion. As a grand general approximation for the hydrocarbon fuels, 

h == 19.6 M.W., a =s 4.5> n s; 0.6 M.W., 
where K.W. is the molecular weight of the fuel. This gives h/na 2: 7.3 and 

IaaI , 

—j- = ■— Jto(8.3) = 0.29 • 
[approximation for all hydrocarbon fuels] 



212 

88 
APPENDIX 3 

Gross Energy Requirement of Nuclear Fuel 

Nuclear fuels may be treated in the same way as fossil fuels. 
That is 

1. 1 kg of nuclear fuel undergoing fission releases a definite 
amount of heat (of which 95$ is prompt). Thus: 

1 kg 233 U( fissioned) = 8.2 x 1Q 13 J 

= 1.96 x 10 10 kcal 

= 2.28 x 10 7 kwh 

( - 1.96 x 10 3 toe ) 

( ^ 1.50 x 10 3 toe ) 

2. The GER should take into account the energy to make the nuclear 
reactor system work and any unfissioned material sent to recycle. 



213 



REFERENCES 



1. Gloyne, A. Keep the Wheels Turning, Science and Technology Society 
Symposium, Stirling, Scotland. November 1973* 



.2. Odum, H.T. Environment, Power and Scoiety, Wiley Interscience, 
New York 1970. 



3. Pimentel, D. et al. Science, November 1973i PP ^3-^9. 

4. Slesser, M. J. Sci. Food Agriculture, 1973, 2h % pp 1193-1207. 



5- Leach, G. Man-Food Equation. Edit. Arthur Bourne, Academic Press, 
due 1975. 



6. Chapman, P. Metals and Materials, February 197^, p 107- 



7. Berry G. and Fels, M. The Energy Cost of Automobiles, Science and 

Public Affairs - Bui. Atomic Scientists. December 1973, 
p 11. 



8. Elbek, B. Consumption of Energy in Denmark and its Rate of Growth: 

Report to I.F.I.A.S. , Stockholm, Sweden. 

9. Herendeen, R. Input-Output matrix for the U.S.A., 1963. C.A.C. Document 

69, Centre for Advanced Computation, Urbana, Illinois 61801, 
U.S.A. 



10. Just, J. Impacts of New Energy Technology using generalised input- 

output analysis 

11. Leach, G. and Slesser, M. Energy Equivalents of Network inputs to 

Agriculture, Strathclyde University, 1973- 

12. Chenery, H. Studies in Structure of the American Economy, 

Edit: W. Leontief, Oxford University Press, 1973. 

13. Long, T.V. The integration of Energy into economic theory, Department 

of Chemistry, University of Chicago, 197*+ • (unpublished report). 

Ik. Herendeen, R. and Bullard, C. Energy cost of goods and services, 

Centre for Advanced Computation, Urbana, Illinois 6l801, U.S.A. 

15. Chapman, P., Leach, G. and Slesser, M. Energy Policy, September 197^ 



214 



IFIAS 

Report no 9 



WORKSHOP REPORT 

INTERNATIONAL FEDERATION 

OF INSTITUTES FOR ADVANCED STUDY 

WORKSHOP ON 
ENERGY ANALYSIS AND ECONOMICS 



LIDINGO, SWEDEN 
JUNE 22-27, 1975 



215 



WORKSHOP REPORT 



INTERNATIONAL FEDERATION OF INSTITUTES FOR ADVANCED STUDY 



WORKSHOP 



ON 



ENERGY ANALYSIS AND ECONOMICS 



LIDING3, SWEDEN 
JUNE 22-27, 1975 



© IFIAS. Stockholm 1975 



216 



The International Federation of Institutes for Advanced 
Study, IFIAS, is a non-governmental, non-profit group of 
some twenty institutes of advanced study in sixteen 
countries. It was established 1972 under the auspices 
of the Nobel and Rockefeller Foundation as a potentially 
significant and new instrument for truly transdisciplinary 
and transnational efforts among physical, biological and 
social sciences and the humanities. 

IFIAS aims at mobilizing not only its own organization but 
also other institutions, scholars and experts to examine, 
with the aid of existing knowledge, some of the increasingly 
complex and interrelated problems of our world. 

This report forms a part of IFIAS project 

"Energy and Quality of Life" 
which is one of IFIAS current activities. The others are 

- Impact of Climate Change on the Character and Quality of 
Human Life 

- Human Settlements: Understanding their Nature and ^uiding 
their Development for the Benefit of Man 

- Socio-economic and Ethical Implications of Enzyme 
Engineering 

- Management and Policy Options for Regions Faced with Water 
Shortage 

- Interaction of Health, Nutrition and Education on Human 
Growth and Development 

- Soil Resources of the Earth, their Utilization and Pre- 
servation 

The Executive Secretariat of IFIAS is situated in Stockholm, 
Sweden, Address: P.O. Box 5344, Postal code S-102 46 
tel. 08 - 631315 



217 



ACKNOWLEDGMENT 



The participants in this meeting and IFIAS are deeply 
indebted to the IBM corporation for the superb hospitality 
extended by the staff at the IBM Nordic Education Center in 
Lidingo, Sweden. Financial support by the Swedish Board for 
Technical Development and the United States National Science 
Foundation Office of Energy Research and Development Policy 
is gratefully acknowledged. 

Additionally, the Workshop members express their gratitude 
to Dr. Sam Nilsson, Mr. Per Lindblom, Ms. Barbara Adams and 
Ms. Els van den Berg of the IFIAS staff for excellent 
operational planning, efficient management, and many individual 
courtesies. Finally, our thanks to Ms. Barbara Tinsman of 
The University of Chicago for typing the final draft of the 
Report. 



68-391 O - 76 - 15 



218 



II 



PARTICIPANTS 



Dr. Frank Alessio 

Criterion Analysis Inc. 

8350 North Central Express Hwy. 

Suite 822 

Dal I as , Texas USA 

Tel: 214-692-1453 

Professor R. Stephen Berry 
Department of Chemistry 
The University of Chicago 
Chicago, Illinois USA 
Tel: 312-753-8286 

Ms. Sussanne Blegaa 
Physics Laboratory i!l 7 Bldg. 309 
Denmark Tech. University 
D-2800 Lyngby, Der.mnrk 
Tel: 02 -881611 

Dr. Peter Chapman 

Energy Research Group 

Open University 

Walton Hall 

Milton Keynes 

MK7 6AA 

England 

Tel: 0908 - 74066 

Dr. J. -P. Charpentier 

IIASA 

Laxenburg 

A 2361 Austria 

Tel: 02236 - 7485 



Professor Willem van Gool 
Analytisch Chemisch Laboratorium 
Croesestraat 77A 
Utrecht, The Netherlands 
Tel: 030 -880611 

Dr. M.K. Hcmid 

Kuwait Institute for Scientific Research 

P.O. Box 12009 

Kuwait 

Tel: 818585 

Dr. Asger Hansen 

Niels Bohr Institute 

Blegdamsvej 17 

D-21 00 Copenhagen, Denmark 

Tel: 01 - 321616 

Dr. William Hogan 

Office of Quantitative Methods 

(Room 4530) 

Federal Energy Administration 

Washington, D.C. 20461 

USA 

Tel: 202-961-8462 

Professor Tokao Hoshi 

College of Er.oineerir.g 

Seikei University 

Kichijoji-kita 

Musashino-ski 

Tokyo 180, Japan 

Tel: 0422 -515181, ext. 599 



Professor Bent Elbek 

Niels Bohr Institute 

Blegdamsvej 17 

D-21 00 Copenhagen, Denmark 

Tel: 03 -356035 



Professor Lawrence Klein 

Economics Department 

Wharton School of Finance end Commerce 

University of Pennsylvania 

Philadelphia, Pennsylvania 19104 

USA 

Tel: 215-243-7713 



219 



III 



Dr. Lars Kristoferson 
Department- of Plasma Physics 
Royal Institute of Technology 
S-l 0044 Stockholm, Sweden 
Tel: 08 - 236520 

Professor Tjalling C. Koopmans 

Department of Economics 

Yale University 

Box 2125 

New Haven, Connecticut USA 

Tel: 203-436-2578 

Mr. Gerald Leach 
3 Tanza 'Road, Hampstead 
London NW3 2UA, England 
Tel: 01 -4359025 

Drf J.M. Leathers 
Expcutlve Vice Preriden* 
Dow Chemical USA 
Barstow Building 
2020 Dow Centre 
Midland, Michigan 48640 
USA 
Tel: 517-636-6852 



Mr. William Martin 

Massachusetts Institute of Technology 

Energy Laboratory (E40 - 159) 

Cambridge, Massachusetts 021 39 

USA 

Tel: 617-235-6824 

Dr. Bert Mclnnis 

Structural Analysis Division 

Statistics Canada 

24th Floor, Coats Blag. 

Tunney's Pasture 

Ottawa, Canada 

Tel: 613-995-0635 

Dr. Sam Nilsson 

IFIAS 

Box 5344 

S-10246 Stockholm, Sweden 

Tel: 08 -631315 

Dr. Peter Roberts 

Department of Environment SI 2/03 

2 Marsham Street 

London SW1, England 

Tel: 01 -2128461 



Dr. Thomas Veach Long, II 

Department of Chemistry 

The University of Chicago 

5735 South Ellis Avenue 

Chicago, Illinois 60637 

USA 

Tel: 312-753-8263/753-8286 

Mr. Amory Lovins 

c/o Friends of the Earth Ltd. 

9 Poland Street 

London W1V 3DG, England 

Tel: 01 -4341684 



Dr. Malcolm Slesser 
Energy Studies Unit 
University of Strcihclyde 
100 Montrose Street 
Glasgow G4, Scotland 
Tel: 041 - 5524400 

Dr. M. R. Srinivasan 
Department of Atomic energy 
Government of India 
Calaga, Bombay 5, India 
Tel: 215229 



220 



IV 



Professor Ingemar Stahl 

Department of Economics 

Lund University 

Fack 

S-22005 Lund 5, Sweden 

Tel: 046 - 124100 

Professor G. Swarup 

Tata Institute of Fundamental Research 

Homi Bhabha Road 

Bombay 400005, India 

Tel: 219111 

Professor Pinhas Zusman 

Development Research Centre, IBRD 

1818 H Street, NW 

Washington, D.C. 20433 

USA 

Tel: 202-477-6174 

Current: Hebrew University 

Faculty of Agriculture 

Rehovot, Israel 



221 



RAPPORTEUR'S NOTE 

This Report is drown from several sources. Participants submitted both 
working papers and published articles for consideration at the Workshop. By 
common cgreement, portions of these were assimilated into the Report without 
attribution. Session summaries were prepared daily by a rotating group of four 
or five. Recordings were made of several sessions, and extensive running notes 
were kept. Lastly, all participants had the opportunity to comment on an initial 
draft, and the majority of their suggestions have been incorporated in the final 
version. 

The Workshop discussions ere not reported in chronologic,;', order, and there 
were only a few items about which there was a unanimity of views. Consequently, 
this Report is a synthesis that inevitcbiy incorporates some of the prejudices of the 
rapporteur. Some of the energy analysts criticized the first draft of the Report for 
what they viewed as a strong emphasis on the incorporation of the data from energy 
analyses into economic decision-making rath* than op it: utilization in c more direct 
evaluative role. I have attempted to meet this criticism in the final draft in order 
to make this statement more fully representative of the opinions voiced at the Workshop. 
However, I am also willing to be counted as a strong supporter of the former view, 
and the reader should be forewarned. 

Ihomas Veach Long, II 



222 



VI 



WORKSHOP SUMMARY 

Twenty-seven economists end scientists from ten countries attended this 
Workshop, which considered the relationships between economic analysis and energy 
analysis. Energy analysis is a new field whose objective is the evaluation of resource 
flows in societal processes using physical units. 

While we agreed that one of the principal roles of energy analysis is to furnish 
information that may be utilized in the allocation of the scarce resource energy, the 
Workshop unanimously concluded that this important function should not be interpreted 
as implying an energy theory of value. Th?s|c<*>'r.clusion'rests orj<the simple observation, 
applicable across a wide range of institutional forms and degrees of technical develop- 
ment, that besides energy resources there are often indispensable primary inputs — labor, 
land, capital, non-energy minerals — with equal claim to having their scarcities (relative 
to the needs for them) expressed in the valuation system that guides allocation. However, 
many participants believe that the results of energy analyses would have their greatest 
impact if they were presented in a form suitable for incorporation into a valuation system. 

Thus, the Workshop stressed the complementarity between economics and energy 
analysis, rather than elements of competition. The economist and the energy analyst 
alike believe that energy analysis can provide important physical data for the economic 
analysis of current productive and consumptive activity. Some Workshop members em- 
phasized the use of energy analysis in assessment of activities in which there is clear 



223 



VII 



evidence of market imperfections or failures. Additionally, energy analysis may be 
useful in technology assessment and in the evaluation of the impact of macroeconomic 
policies on energy demand. Technological assessments of this kind can also be used to 
define constraints within which a viable economic society must exist. It was clear That 
the interest of some energy analysts is not only with framing a response to scarcity of 
energy resources, but also with evaluating long-term destructive effect? of intensive 
energy use. 



224 



VIII 



TABLE OF CONTENTS 

Section Page 

Acknowledgment I 

List of Participants II 

Rapporteur's Note V 

Summary V I 

Table of Contents VIII 

List of Figures IX 

Introduction 1 

What Is Energy Analysis? 3 

Energy Analysis and Economics 6 

General Objectives and Opening Forays 12 

Valuation and an Energy Theory of Value 20 

Economic and Physical Efficiency Criteria 25 

Engineering and Economic Production Functions 34 

Optimization Over Time 43 

Setting the Limits 49 

The Economists' Critique of Energy Analysis 57 

The Energy Analysts' Critique of Economics 60 

Economics-Energy Analysis Interfaces 63 

Issues for Further Thought 65 

Reports of Empirical Studies 67 

Appendix I. Guidelines for Energy Analysis 77 

Appendix II. A Thermodynamics Primer 93 

Bibliography 99 



225 



IX 



LIST OF FIGURES 

Page 

Figure 1. Two extrapolations of past trends in world fuel consumption. 17 

Figure 2. The convex hull for the production of electrical energy and 

equipment. 29 

Figure 3. Production with two primary factors, energy and labor. 33 

Figure 4. The material and fuel inputs to a steel furnace. 40 

Figure 5. Integrated production of automobiles and steel. 42 

Figure 6. Gross energy requirements for ammonia production. 51 

Figure 7. Distribution of energy use per capita per year (1971) for 72 
178 countries. 

Figure 8. Levels in the definition of the system boundary. 89 

Figure 9. Symbols for energy flow diagrams. 90 

Figure 10. Typical flow diagram for aluminum production. 91 



226 



INTRODUCTION 

The Report of this Workshop will be interesting to a diverse audience: to 
the economist and the energy analyst primarily, but also to the politician, the 
governmental decision-maker, and the industrial manager. Others who will find 
the Workshop results intriguing are those who are concerned with intellectual inter- 
faces. Through confrontation and resolution, criticism and reasoned reply, similarities 
and contrasts in goals and methodologies were delineated, and new insights were 
formed. By one participant's criterion, the purpose of such an endeavor is to change 
the way people think. On this basis, the Workshop was a success. There is little 
question that both the energy analyst and the economist left Sweden with remarkably 
modified views. Many of us feel that rumination over the issues that were raised will 
continue to produce substantive results. 

This was the second Workshop on the subject of energy analysis that has been 
sponsored by IFIAS. The first Workshop was held in Guldsmedshyttan, Sweden during 
August, 1974. At that meeting, the methodological problems of this young field were 
discussed, and a set of procedural recommendations were formulated. A summary of 
the recommendations is provided in Appendix I. Members of the first Workshop recognized 
that the interface between energy analysis and economics should be examined, and 
they recommended to IFIAS that a workshop on this subject be held during the following 
summer. 

There were twenty-seven participants in the second Workshop, assembled at 
the IBM Nordic Education Centre on the Lidingo peninsula outside of Stockholm during 



227 



the last week of June, 1975. The members of the Workshop represented twelve countries. 
Included in this group were several of the earliest practitioners of energy analysis, 
a world-renowned econometric ian 7 scientists and economists directly involved in the 
determination of national and international energy policies, academic economists and 
scientists, and an executive with a major industrial concern that has employed energy 
analysis in management for a number of years. We were particularly fortunate to count 
as a member of our group one of the 1975 Nobel Laureates in Economics. This diversity 
in backgrounds was evident throughout the proceedings. 



228 



WHAT IS ENERGY ANALYSIS? 

Broadly, energy analysis is a field devoted to studying societal use of a 
single aggregate resource, energy. Usually we think of energy as being provided 
only by fuels or by renewable sources such as solar, wind or hydro power generation. 
However, thermodynamics tells us that all materials have a potential for furnishing 
energy. This is even true of those that are not ordinarily considered to be fuels. The 
material flows in a process have associated with them flows of thermodynamic, potential 
to do work. 

Energy analyses quantitatively trace the changes in the thermodynamic potentials 
of materials as they pass through successive process stages. In productive processes, the 
thermodynamic potentials of materials often increase. This is because energy has been 
added to the materials through the application of heat energy from electricity or fuels 
or by doing work on the system. However, part of the heat and work energy is inevitably 
lost in the transfer process. Thermodynamic laws indicate that there are inviolable limits 
on the physical efficiencies of energy transfer processes. 

There is an additional consideration. Heat and work that are added to the 
system in one part of a process may be lost quasi -simultaneously in another part of the 
process. Indeed, for a total process, the thermodynamic potential (stored energy ) of 
the materials may actually decrease, even though there were additions of heat and work 
from external sources. 

An example of this would be a driven cxoihermic chemical rcccHon process. 
An exothermic reaction is one that gives off heat, and the reactants have a higher 



229 



thermodynamic potential than do the products. However, one can make the reaction 
go faster by raising the temperature of the system through the addition of heat. 
Industrial processes that use fuel and electrical energy but which result in a reduction 
in the thermodynamic potential of the materials are not uncommon. 

One goal of energy analysis i? to indicate where reductions in the energy 
requirements for total processes could be made - the pressure points for technological 
change. Possible reductions are assessed by first quantitatively evaluating the actual 
energy furnished to the process in the form of fuels and electricity. This is then 
compared with the actual change in the thermodynamic potential of the materia!. The 
difference in these two quantities is the energy that is lost in the process. 

Thus, there are two quite different senses in which an energy analyst sees energy 
as being "embodied" in a material. One sense is much the same as that of an economist 
when he speaks of "embodied labor. " In this accounting sense, direct fuel and electrical 
energy that are utilized in the process form a portion of the "embodied energy. " 
Additionally, the total fuel and electrical requirements for producing the inputs, tracing 
back to raw material extraction, are the indirect energy inpuis. 

The second sense in which energy is "embodied" in the materials is as thermodynamic 
potential. A brief discussion of a few thermodynamic concepts is given in Appendix II. 
"Embodied energy" in this sense can be used to do work and has a real physical meaning. 
Energy analysis focuses on the discrepancy between this real change in thermodynamic 
potential in a process and the "embodied energy" requirements calculated by the 



230 



accounting procedure. Note that a slippage loss of materials is captured by both 
evaluations. Thus, an energy analysis superimposed on a materials flow incorporates 
information about both energy and materials flows. 



231 



ENERGY ANALYSIS AND ECONOMICS 

It may be helpful to put forward a few observations and to do away with some 
misconceptions in order that subsequent meetings do not have to pay the heavy search 
costs associated with finding a basis for interaction between these two fields of analysis. 
First, the common ground between energy analysis end economics is found in their 
parallel claims to be (in part) sciences of description. The motivating force behind the 
initial energy analyses was an attempt to construct accurate and all -encompassing 
descriptions of production and consumption processes. For example, there was the desire 
to account for externalities. These are not fully captured by operation of the market 
by definition. Thus, it is no accident that o majority of the ecrly worker? in energy analysis 
are physicists, physical chemists, and ecologists. These disciplines concentrate on 
precise specification as a basis for scientific progress. 

Unfortunately, this primary goal has been obscured by the emphasis on the use 
of energy analysis in framing responses to perceived energy supply constraints. 
The concentration on the role of energy analysis in policy decisions has led economists 
to think of energy analysis as though it were a competing methodology for determining 
the efficient allocation of scarce resources over space and time, which is the focus of 
economics. Indeed, a few energy analysts apparently see this as its role, and suggestions 
of an "energy theory of value" have peppered discussions of the subject. It was the 
unanimous view of the participants that a value system based on the single factor energy 



232 



is not satisfactory for analyzing modern mcrket, mixed, or planned economies (vide infra). 
This conclusion rests on a simple observation applicable across a wide range of institutional 
forms and degrees of technical development. Besides energy resources there are other 
indispensable primary inputs - labor, land, capital, non-energy minerals - with equal 
claim to having their scarcities (relative to needs for them) expressed in the valuation 
system that guides allocation. For this reason, it is necessary that the needed precise 
description of technological processes not be limited to the energy inputs, but include 
all important inputs and outputs. The principal goal of energy analysis is the develop- 
ment of a portion of the precise physical description of the operation of real-world 
processes. This description does not supplant that of economic analysis, but supports 
and complemen ts it and may provide new perspectives. 

This should not be taken to diminish the role of energy analysis in developing 
policy alternatives. To the contrary, energy analysts can concentrate on determining 
those situations in which the physical description may provide a useful addition 
to market information. A variety of questions of this type can only be answered 
empirically. For example, are there circumstance:; under which energy analysis furnishes 
faster (and equally accurate) signals of impending critical situations than does the market? 
Operationally, under what conditions can one accept the reduced information content 
of energy analysis as compared to economic analysis, because the costs of carrying out 
the energy analyses are also smaller? Economic analyses may also require more time 
than energy analyses. Consequently, one may be prepared to sacrifice the more complete 



233 



information incorporated in the economic data. A plant manager may employ energy 
analysis as a materials control technique because it furnishes him v/ith a near-instan- 
taneous picture of his operation that is not subject to market fluctuations. Economic 
analysis, which should be used in parallel, requires a time lag for formulation and is 
sensitive to variation in prices. These are questions of comparative sensitivity, cccuracy, 
time responses and ease of informational organization that can be answered only by 
careful empirical evaluations. 

Second, the descriptive framework of the energy analyst and the economist have 
different bases, and this can lead to thorny misconceptions. Traditionally, the economist 
has focused on inputs and outputs, regarding the transformation process from the former 
to the latter as a black box. However there is substantial interest in incorporating 
more accurate process description, particularly among econometric ians and other modelers. 
As one economist pointed out, energy and materials are washed out in macro-econometric 
models in order to avoid double counting. Only capital and labor are included as 
original productive factors, although land may also appear for the case of a frontier 
economy. Energy and materials may be incorporated in models of smaller scope. Con- 
versely, the energy analyst trained in thermodynamics (see Appendix II) concerns himself 
with the transformation process and the factors that effect the change. From this view- 
point, energy is of prime importance because every process requires a change in energy 
(free energy, more precisely). 



i - 3 « > 1 O - 7fi 



234 



These are radically different frameworks and should be so recopnized. The 
central question is whether energy is an intermediate good? This is answered affirma- 
tively by the econometric ian, acting from a methodological posture, and negatively 
by the energy analyst, whose different philosophical framework requires him to consider 
it a primary descriptive element. The economist is interested in fuels and feedstocks, 
and the energy analyst is interested in energy. As will be discussed below, a bridge 
between these two conceptual foundations may possibly be found in some economic 
literature dealing with engineering production functions. 

A third point that came across strongly was that neither economics nor energy 
analysis are monoliths. Within each of these disciplines there is a broad range of points 
of view. In economics, one can identify the econometrician, whose outlook is partially 
determined by what is possibly measurable and what is required to measure it; the 
analytical economist, who concerns himself with the construction of the conceptual 
framework; and at least three forms of political economists, one sort emphasizing the 
capacity of a freely-operating market to compute efficient solutions and to implement 
them, another more interested in questions of distribution, and the third concentrating 
on the operation of planned economies. Of course, a number of subdivisions could be 
added to this abbreviated list. 

Despite the short history of energy analysis, there are sharp contrasts between 
at least two schools of thought. The first argues that energy analyses stand alone, 
totally independent of economic analyses, and that choices can be made on the basis 



235 



10 



of either or both. This position emphasizes the possibility of carrying out any desired 
transformation of a material resource if sufficient energy is available. Energy should 
therefore be treated as a unique, essential resource. The second school basically 
contends that information from energy analyses can and should be incorporated into 
economics. This school tends to see energy as but one of the primary natural resources 
that dominate technological description. Water would be another. They suggest that 
economic decision-making can be improved by capturing all the physical information 
that it can. Of course, most energy analysts find themselves somewhere in the middle 
between these two cases. 

Another important view of the relationship between energy analysis and economics 
transcends the other two positions. This is the idea that economic processes must 
operate within the constraints imposed by me physical" and biological world. Techno- 
logical descriptions like energy analyses are useful in that they permit quantitative 
assessments of the constraints. Thus proposals should be subjected to two assessment 
stages. First, viability should be determined on purely physical grounds. Then, value 
and possible tradeoffs can be analyzed using economic criteria. 

A fourth observation is that the interface being explored is between two fields 
at vastly different stages of maturity. Economics has a long and largely successful 
history of being a useful tool in marshalling the potentials of society into productive 
activity. Energy analysis is a much newer endeavor and has little to point to in the 
way of historical contributions. There are many questions that energy analysis has 



236 



11 



not had time to answer, but this does not mean that they will not be c cannot be 
answered in quite a robust way. It is a field for which the data base is not yet 
sufficient to allow a response to questions of its utility, such as "what can energy 
analysis do that economics doesn't do better? " Some felt that there is a possibility 
that energy analysis will not survive as a well-defined discipline. In view of this, 
it would have been all to easy for the economists present to criticize energy analysis 
soundly and then to dismiss it as a rudimentary exercise in tracing the flow of a single 
resource through economic society. To their credit, they did not, but instead, through 
gentle persuasion, contributed measurably to the growth of this young field. 



237 



12 



GENERAL OBJECTIVES AND OPENING FORAYS 

The Workshop was organized about three objectives, to examine: 
1. energy analysis as a complement to financial accounting for evaluating 

new and alternative technologies; 
2. energy efficiency versus economic efficiency as a criterion for resource 

allocation; and 
3. the integration of physical information into economic behavioral relation- 
ships. 
These objectives were quickly broadened when energy analysts suggested that the 
proper role of energy analysis is in evaluating the energy implications of policies, 
but not as the sole decision basis for determining policies. Further, they suggested 
that potential evaluative functions of energy analysis could be broken up into those 
appropriate for short-term, medium-term, and medium-long-term decisions. 
Short-term: 

a) calculation of fuel price elasticities; 

b) evaluation of "energy conservation" measures; 

c) testing production function specification and some price system assumptions. 
Medium-term: 

d) disaggregated demand forecasting; 

e) evaluation of alternative energy sources. 



238 



13 



Medium-long-term: 

f) documentation of one effect of resource depletion; 

g) prediction of when the costs of production will rise based on technical factors; 
the inclusion of time scales and non-linearities; 

h) description of "points of futility " associated with "technological (physical 
and biological) limits, " and placing limits on allocation over finite times; 

i) setting limits on ex ante production functions; 

j) analyzing the stability of societal trajectories based on physical resource 
use. 

A participant then outlined energy analysis from an economist's viewpoint. 
Several features of the outline formed a basis for subsequent discussions. It is worth- 
while reconstructing the presentation in some detail because it is representative of an 
economist's perspective. 

The economist asserted that the existence of energy analysis as a separate meth- 
odology is based on four propositions and that the burden of their proof rests with the energy 
analyst: 

1 . all energy resources are scarce; 

2. this scarcity increases over time; 

3. scarcity imperils the quality of life; and 

4. society must focus on this scarcity by employing criteria of physical efficiency. 



239 



14 



In order to respond to this perceived scarcity situation, the economist's 
impression is that energy analysis has developed two analytical objectives. One 
analytical objective is the determination of physical efficiencies at a micro level, 
often in a'single process, in order to assess the state of technology. The second is 
the establishment of thermodynamic boundary conditions for these processes with the 
intent. of defining the limits for energy husbandry efforts. Also, some energy analysts 
have championed a "methodological tour de force, " which economists clearly reject, 
' in introducing an "energy theory of value. " A good deal of discussion was devoted 
to this in subsequent sessions, and it will be dissected below. 

Insofar as potential relationships with economic analysis are concerned, it 
appears that energy analysis could aid in identifying alternatives in responding to price 
changes, thereby helping to minimize search and information costs. More importantly, 
in the absence of well-formulated future plans or futures markets, energy analysis could 
aid by providing accurate imputed prices faster. The economist, drawing on analogues 
in his own field, quickly identifies a list of methodological problems. These include 
those of defining the system boundary,of the need for some assumption regarding the 
specification and aggregation of heterogeneous inputs such as the fuel mix in the system , 
and of incorporating values and prices into the analytical system. The economist 
approaches energy analysis with the attiude that the energy analyst must show when 
joules can be more useful than constant dollars in analyzing the decisions of economic 
society. 



240 



15 



Some of the energy analysts sharply disagree with the assertion that the only 
use for their discipline is in response to the scarcity of energy resources. Their view 
is that the economist's "scarcity" implies only those conditions required in the responses 
of a market. By contrast, energy analysis, in furnishing a scientific basis for energy 
husbandry, is equally well a diagnostic tool for indicating overly intensive energy use 
relative to physical constraints, even when these constraints are not yet associated with 
costs. An energy analyst concerned with intensive energy use presented the following 
example. 

To a first approximation, the only way the earth can increase the amount of 

heat it radiates into space is through an increase in its surface temperature. According 

to the Stefan-Boltzmann radiation law, the annual solar energy input, E, is related to 

the surface temperature, T, by 

E = a AT 4 

where 

a = Stefan-Boltzmann constant 
A = radiating surface area. 

In thermal equilibrium this is also equal to the energy output into space. 

18 
Using an annual solar input of 3.6 x 10 MJ/annum this formula gives a surface 

temperature of 280°K, which is reasonably close to that which is observed. Working 

from this equation, we can also derive a relationship between the change in energy 



241 



16 



input and change in surface temperature, namely 

dE _ . 61 
T " 4 T 

Thus if the energy input were increased by 1%, a net addition of 3.6 x 10 MJ/annum, 
then the rise in surface temperature would be 1/4% or 0. 7°C. This could be serious, be- 
cause a change of this magnitude could lead to a melting of the polar ice-caps. This 
level of fuel consumption is within our time-horizon. Two extrapolations of world fuel 
consumption are shown in Figure 1 . That marked (b) assumes a constant growth of 
5% p. a, the present growth rate. That marked (a) assumes a continued increase in the 
rate of growth. Both put the 1% solar level within our own time horizons, This 
example indicates why some energy analysts are concerned with abundance rather than 
scarcity of energy resources. 

Moreover, scientists generally agree that local climatic effects will sound the 
alarm bells before the global problems arise. For example, according to some reports, 
London annually dissipates 20% of its annual solar input. The temperature in the middle 
of London is about 5°C higher in summer and 10 C C higher in winter than its surroundings, 
which could lead to secondary local climatic disturbances. 

This discussion is intended-io.jpe i I lustrqtive, fa/her than definitive. We do not 
intend for the values of parameters cited to be taken as precisely true, although the 
magnitudes are at least crudely correct. This example indicates that the data from 
energy analysis on waste heat production might be used to establish the physical constraints 



242 



17 



3.6xl0 17 

3 

« 3.6x10 

CL 



15 



g ' 3.6x10 



3 3.6xl0 14 
Z 

8 



3 13 

"- 3.6xlO ,J 
Q 



3.6x10 



12 





1 

1 % of solar input 






/ 


/ 










, 














/ 


/ 












(a)/ 


/(b) 












//' 














/ 












/ 














































!_ 


j »._•>» im'wcri^ 



I860 1900 1940 1980 2020 2C60 2100 

YEAR 



Figure 1: Two extrapolations of past trends in world fuel consumption. 



243 



18 



within which economic systems must operate. Here is one case where these arise 
from intensive energy use rather than from the constraint of energy scarcity. 

There is another concern of energy analysis that is somewhat separable from 
considerations of energy resource scarcity in the first order. Energy analysis can be 
used as one indicator of what was termed by a Workshop member "the saturation of 
technical progress. " Particularly in the mineral extraction industries and in agriculture, 
increased fuel inputs may be required simply to maintain present production levels. 
Decreasing soil qualify under intensive farming and the need to use even lower grades 
of mineral resources could lead to increasing requirements for all factors of production - 
capital and labor as well as energy. 

The historical experience in many of the extractive processes is that more 
intensive use of items of physical capital ha* resulted in ded leasing requirements for 
both labor and energy inputs per unit output. However, there are suggestions that ihis 
trend has reversed in the case of energy use, and energy and capital may now be 
complementary factors in some industries. 

The argument is that resource grades are decreasing at a rote greater than the 
rate of factor-saving technological innovation in these processes. Furthermore, the 
argument implies that energy analysis can be a sensitive indicator of this effect. 

Of course, increasing efficiency in the use of capital and labor could generate 
greater productivity even in the face of the larger fuel requirements if "technical 
saturation" did not lead to increased requirements for all factors. Technical saturation 



244 



19 



in the use of energy alone is important only to the degree that energy is scarce. 

This reflects the long-range concern of many energy analysts with the more 
general question of how society uses all of its natural resources, including but not 
solely restricted to scarce energy resources. The degree to which energy analysis 
can furnish information about technological change is a question that must be 
investigated empirically. 

Although the energy analysts did not agree with several portions of the economist s 
comments, it was clear that in order to establish an initial basis for communication, 
the problem of valuation had to be considered. The "energy theory of value" had to be 
dealt with, if only to dismiss it. 



245 



20 



VALUATION AND AN ENERGY THEORY OF VALUE 

There are two senses in which the term "valuation" is generally employed. 
The first is its use in a normative connotation as a synonym for a purposeful optimi- 
zation that is imputed to society, such as the maximizction of a social welfare function 
or the definition of efficient growth paths over time. Another sense of the term is as 
a description of the preferences of consumers as revealed in consumer demand in the 
market. The economists attending the Workshop felt that in order for energy analysis 
to become something, to realize its full-potential, an element of valuation must be 
introduced into its structure. In order to do this, it is necessary to adopt one of three 
posiures. One choice would be to minimize energy use as the valuation procedure. 
A second is to recognize the existence of other scarce factors but to base their 
evaluations on some measure of "embodied energy" only, and cgain minimize energy 
use. Finally, factors of production such as capital and labor could be entered and the opti- 
mization could be based on their scarcities as well, as is ordinarily done in economic cnalysis 

A few energy analysts have ciecrly pur themselves en record as favoring a 
method of valuation based on energy minimization alone. Hannon has stated: 

"The adoption of a national — and consequently a personal — energy budget 
appears to be necessary. . . Individual allocation could be similar to that of 
our present economies, which reflect'personal value, except that we would 
have to strive for the right tc consume energy; the accrued currency would 



246 



21 



be regulated by the amount of energy budgeted for a given period. . . 
Recognition of the value of energy is equivalent to setting energy as the 
basis or standard of value. In doing so, society readmits itself into the 
natural system in which acknowledgment of energy's importance has never 
been lost. " } 
In a like vein, Odum suggests that a large number of economic concepts ( e.g ., goods, 
wants, income, interest, inflation) are best analyzed in an energy language in which 

these terms are redefined. He uses these redefinitions to support a "technocratic" 

2 3 

value system with an "energy certificate" as a "money standard. " ' He postulates 

that money flows and energy flows circulate in opposite directions, and this leads him to 
assume that a functional relationship exists between these flows. "Money is a counter 
current with the ratio of money to energy flow being price. " The logic establishing this 
relationship is unclear, and he has unfortunately eschewed the use of conventional 
thermodynamics in energy analysis in favor of a system that seemingly directly in- 
corporates an efficiency concept. 

There are several superficial attractions to a value system based on a 
single factor. The system is holistic, and its predictive value is substantial. As 
Samuelson has pointed out regarding the one -factor hypothesis: "A spy can memorize 
(the technological coefficients) and know most of what there is to know about the 
economy. " But the shortcomings are also clarified. Energy analysts will already have 
realized that their methodology leaps right across production to consumption without 



247 



22 



mentioning a market connecting the two. Indeed, for a long-run one-factor Ricardian 
economy, the "substitutability theorem" can be shown to hold, which states that even 
though possibilities for substitutions exist, no substitutability need be experienced. 
In such a world, barring joint production, relative prices are determined by techno- 
logical coefficients and are independent of the mix of consumer demand. Thus, complete 
economic control is implied, whatever level of disaggregated decision-making is adopted. 

The market economist has confronted single factor theories of value such as 
those based on land or on labor before. He has refected them because they are weak 
in treating processes that involve more than one factor - in particular, processes that 
take time and therefore involve the participation of capital. The economists who were 
members of the Workshop felt that an energy theory of value was a poorer approximation 
to reality than the analytically similar labor theory of value. This is because, in addition 
to neglecting to treat time properly, there is no scarcity value ascribed to human labor. 
The energy analysts accepted this position. Thus, the Workshop rejected the concept 
of on energy standard of value — not because it is impossible to design such an allocation 
mechanism, but because such a system does not adequately describe the full texture 
within which human economic choices are made. 

Before departing from the topic of valuation, perhaps we should take note of a 
distinction that may be significant in understanding this interface between energy 
analysis and economics. The distinction is between the use of the words 'choice 1 and 
'decision. ' Economics is concerned with the allocation of scarce productive resources 



248 



23 



among alternative uses either now or in the future. Allocation is a social phenomenon, 
involving the actions of groups of individuals. As viewed by neoclassical economic 
theory, the individual is required to choose but not to decide. The differentiation lies 
in allowing choice to admit possible noncognitive elements. Choice views alternatives 
through a preference function. A decision is the result of resolution of a set of alter- 
native actions by purely cognitive selection processes. The information from energy 
analysis can be used to add to the rational component of a choice or alone as a basis 
for decision. 

The way society values scarce resources, such as energy, capital, labor and 
land, results from social preferences and associated social choice. Results from energy 
analyses can serve as the basis for some kinds of decisions and can aid in the formulation 
of cognitive selection criteria. But their use in social choice is limited to furnishing 
information for that portion of the choice that proceeds by a cognitive mechanism. 

The economists maintain that energy analysis can have maximum impact if the 
analyses are carried out so that the data can be directly incorporated into a valuation 
system, whether in terms of competitive market prices or of imputed and computed shadow 
prices. Their position is that the embodiment idea fades and is not pertinent in describing 
price formation if possibilities of substitution of different primary factors of production 
are present. In opposition, many of the energy analysts argue that the most effective 
use of their analyses is nor through incorporation into a valuation system, but as a direct 
input in policy assessment of societal choices. This assessment can then be compared 
and contrasted with the economic analysis that results from a valuation procedure. 



249 



24 



The energy analysts maintain that their assessments can be used without 
valuation to reject technically non-feasible options. Also, should a mode of action 
seriously threaten the survival of mankind, technical assessments need not be subjected 
to further valuation procedures. For example, many would argue that clear and present 
danger to human health from ozone depletion by fluorocarbon spray propel lants exists. 
They would contend that a ban on the manufacture and sale of spray products should be 
implemented on the bcsis of ihis assessment alone, without recourse 1o a further (economic) 
valuation procedure. In a sense, of course, this argues that a very high price should 
be imputed to use of these items. 

Again, if a positive choice is to be made, we concluded that this can only be 
done through a valuation procedure that incorporates all indispensable primary inputs, 
including labor, land, capital, non-energy minerals, and precisely-analyzed energy 
requirements. 



68-391 O - 76 - 17 



250 



25 



ECONOMIC AND PHYSICAL EFFICIENCY CRITERIA 

A second step in constructing a working relationship between the two fields 
is a clear-cut comparison of the economist's concept of efficient allocation of resources 
and the energy analyst's definition of an efficiency measure. There are a number of 
possible physical efficiency criteria that could be adopted. In Appendix II, definitions 
of both first and second law efficiencies are provided. Another possible choice is 
based on the waste factor described in "Guidelines for Energy Analysis" (Appendix I). 

More important than the particular choice of parameter is the rationale behind 
defining a physical criterion. A point that has not been sufficiently emphasized is 
that the first step in carrying out an energy analysis is the construction of a detailed 
picture of the mass flows in the real-world system chosen for study. This is, in principle, 
a complete materials balance. The material output of the process will require amounts 
of material inputs that are greater than would be required in the ideal case because of 
loss of material as "waste" in the course of production. One could use the analysis to 
compare the actual material input requirements for a given output with those computed 
for the limiting case in which ihere is no slippage. 

For a chemical process, the ideal input requirements of materials and energy 
can be evaluated from the stoic biometry of the reaction provided only that the reaction 
goes to completion. Those who are not trained in thermodynamics may find it surprising 
that energy has entered the picture even at this early stage. There has been no mention 
of fuels, nor need there be to this point. As discussed in an introductory section, each 



251 



26 



chemical species has a thermodynamic potential relative to other species, an 
energetic level that is physically determined. When two species of differing 
potential are brought into contact, they may react spontaneously. If heat is gen- 
erated in this reaction, it may be used to do work, although one would not ordinarily 
consider either reacting species to be a fuel. For example, condensation of mono- 
meric units into a polymer is one such heat-producing (exothermic) process. In this 
process, there is a reduction in the thermodynamic potential of the system associated v 
with the generation and loss of heat. 

Alternatively, another class of reactions may proceed only with the application 
of heat or electricity, raising the thermodynamic potential of the system. Purely 
mechanical processes are also describable using thermodynamics, but ideally they 
often involve little change in the thermodynamic potential of the materials. In 
summary, energy in the form of thermodynamic potential is already embodied in 
materials prior to their involvement in any production step, and it can be used to do work. 

After the materials flows have been traced, the energy required in every process 
step is evaluated. It is important to distinguish between two different evaluations. 
One assesses the actual energy requiremenls for the system, with material slippage, 
and the other calculates the requirements for a hypothetical system operating at maximum 
thermodynamic efficiency with no material slippage. The analysis of the actual energy 
requirement for the process generally is carried out by determining from process data 
the material, fuel and electricity requirements for every step, following each material input 
back to its natural state. 



252 



27 



We can assess the ideal energy requirements only if we first designate the 
constraints associated with the process. In traditional thermodynamics, there is no 
time constraint, so the ideal energy requirements are always evaluated for a reversible 
process (see Appendix II). The reversible process requires an infinite time, and the 
actual system will use extra thermodynamic potential in order to proceed at a finite 
rate. The thermodynamic requirements for this ideal process are calculated for the 
system in which there is no material slippage. Finally one compares the actual and 
the ideal by computing some sort of efficiency parameter. 

We note that an energy analysis fits like a veil over the materials flow diagram. 
Because each material has its own gross energy requirement (which would be the thermo- 
dynamic potential of the material if the subprocess producing that material operated 
ideally), slippage of any component will result in a larger energy requirement as compared 
to the ideal system. Thus, the increase in energy requirements due to materials losses 
will be proportional to the energy requirements of the components that are used 
inefficiently. 

Therefore, the energy analyst's definition encompasses somewhat more than 
simply an evaluation of the efficiency with which the single resource energy is used. 
Achieving maximum physical efficiency requires both th e efficient application of 
energy to the system and the careful husbandry of materials throughout, so that none 
are wasted. Energy analyses are sensitive to both sources of inefficiency and incorporate 
two different forms of information. 



253 



28 



The energy analyst maintains that his efficiency criteria are appropriate 
for assessing a trade-off in the physical world, that there is a natural valuation 
system operating in that world, but that it is only part of the larger realm of human 
activity. What it does not permit — and this is where economics enters — is an 
examination of how systems that require human labor and an investment in capital 
can efficiently combine these resources with those of the physical system. 

Economic efficiency is attained if given resources (including capital, labor 
and natural resources) are combined in such a manner that a higher output of any 
desired good could be obtained only at the cost of a lesser output of some other desired 
good . 

Let us investigate this concept in greater detail by analyzing Figure 2. 
The economy represented in this figure consists of two industries, one engaged in the 
production of electrical energy and the other in the production of equipment. Each 
industry makes only one homogeneous product, and quantities of production are represented 
by coordinates y. and y^, respectively. Both industries utilize labor. For simplicity, 
but without loss of generality, we can assume that one unit of labor input is required 
for one unit of output of electricity or of equipment. Thus, we will omit the third 
coordinate corresponding to labor and utilize a two-dimensional diagram. 

Each point in the plot corresponds to a production technology available to the 
industry. A positive value for a coordinate indicates that the commodity is an output 
of the industry, while a negative value signifies that the industry utilizes the commodity 



254 



29 



* ELECTRICITY 




EQUIPMENT 

»y 2 



Figure 2: The convex hull for the production of electrical energy and equip- 
ment (arbitrary units) with labor as a primary input (y = -1). 

o 



255 



30 



as an input. The electrical power generation industry requires equipment and labor 
inputs and furnishes a net output of electrical energy, while the equipment industry 
requires electricity and labor to produce a net output of machinery. The electrical 
power generation industry can adopt methods A., B., C., and others indicated by 
points in the upper left-hand quadrant. The equipment industry has techniques rep- 
resented by points A~, B 9 , C 9 , . . . available to it. There is no joint production. The 
possibility that some labor will not be utilized can be handled by admitting a poin* 
at the origin Oof the coordinate system, for which one unit of labor is expended with- 
out production of outputs. 

The polygon created by connecting the points A., C., O, C«, B_ bounds the 
production possibilities, and every point on the boundary or inside can be achieved 
by the proper combination of techniques. If there are no exogenous sources of supply 
of either commodity, the only attainable points utilizing one unit of labor will lie 
within the triangle L. OL«. A point in the set of attainable points is efficient if there 
exists no attainable point that is superior in providing greater output of one commodity 
without diminishing the output of the other. The line segment L. L„ is the efficient 
set. Point a is not efficient because the output of both electricity and equip- 
ment can be increased within the attainable set, but point is clearly on the efficient 
boundary. Consequently, for a two-industry input-output model, we have arrived at 
the set of possible combinations of techniques that represent efficient use of labor in 
production. These are combinations of only two methods, A. and B,~. 



256 

31 



Observe that it would have been feasible to produce equipment using technique 
A« with a decrease in electrical energy requirements (per worker) and that electricity 
production could employ method C. , with a higher net energy output (per worker). 
But the set of attainable points along A„ D. correspond to a set of technique combina- 
tions that is everywhere inferior to those of L. L,. One function energy analysis can 
serve is to aid in the development of a technique for equipment production that utilizes 
one unit of labor but less energy per unit produced, as would be represented by point 
D«. The efficient set would then fall along a line connecting A 1 and D^. 

One source of difficulty that the economist and energy analyst encounter in 
seeking a level of discussion is that, in loose terms, the economist concerns himself 
with fuels as intermediate goods and does not recognize energy in the abstract as a 
good. The energy analyst treats energy as an aggregate quantity that is a primary factor 
of production. More precisely, the economist deals only with specific forms of energy 
considered at points in the chain of extraction or interception and processing at which 
an option in the extraction, conversion or utilization exists and may be exercised. This 
concern therefore encompasses a number of scarce primary energy sources such as uranium, 
oil, coal in the crust of the earth or elevated water. This disagreement is meaningful 
to the degree that sources of thermodynamic potential other than those ordinarily regarded 
as fuels are utilized by economic society for their ability to deliver this potential. 
For example, materials whose marketplace values are primarily determined by their 
structural properties or by their ability to provide other services desired by society 



257 



32 



are also sources of thermodynamic potential. The energy analyst is arguing that he 
is providing the information that society requires to make a knowledgable choice 
between use of a material for the energy it naturally embodies and its use based on 
some other characteristic. Through careful empirical evaluations, he is showing 
society the full range of options that it confronts. 

Consequently, let us consider a two-industry economy in which there are 
two primary factors, energy (thermodynamic potential) and labor, each available in 
a given amount. The industries will be taken to be metal mining and equipment 
production. Each requires both primary factors. This economy can be analyzed with 
the aid of Figure 3. First, we will ignore any restriction on labor and make the 
assumptions utilized in discussing Figure 2. The attainable point set is defined by 
the two processes A. and A~ and is L. OL ? . Similarly, the attainable point set 
resulting from ignoring the restriction on energy, M 1 OM~,is defined by the pair of 
techniques B. and B«. Taking both restrictions into account, the attainable point 
set consists of the quadrilateral OL DM«, and the efficient production set lies along 
the two line segments L. D and DM~. Thus either of the two pairs cf methods can be 
utilized in efficient production,and the efficient choice" between 'he pairs depends 
on whether labor or energy is the limiting primary factor. Energy analysis can be 
employed in determining the points A. and A«. 



258 



33 



METAL ORE 



B. and IL : y. = -1 ; no energy 




A, and A 2 : y_ = -1 ; no labor restriction. 



restriction. 



EQUIPMENT 



Figure 3: Production with two primary factors, energy (y_) and labor (y .). 



259 



34 



ENGINEERING AND ECONOMIC PRODUCTION FUNCTIONS 

There are a few serious attempts to explore the manner in which more compre- 
hensive physical information can be introduced into economic behavioral relation- 

8-18 
ships by utilizing a production function structure. A few wordsabout production 

functions are in order. The production function expresses the technological possibilities 

relating outputs and inputs that are faced by a productive unit. The productive unit 

considered may be an aggregate entity, such as a nation, and in that case the macro 

production function relationship is 

Q = f(A,B,...), 

in which Q is the cutput quantity expressed as a flow and A, B, . . . are input quantities, 

again expressed as flows. For the macro function, the inputs are taken to be aggregates, 

such as flows of capital, K, and labor, L. 

If the productive unit is an industry, a firm, a division or a process unit, the 

production function can be thought of as a microeconomic relationship, again expressing 

the technological possibilities confronting the micro unit. The output is now thought 

of as a particular commodity flow,and the inputs are also flows of particular goods and 

services. One can also differentiate between long-run functions,in which all inputs 

can be modified in amount,and short-run functions, in which items of physical capital 

are taken as fixed. A production function formulation presumes that the output is the 

maximum possible from the group of input factors or, equivalently, that a given output 

.19 
is produced by a minimum quantity of inputs — a "premaximization" presumption. 

Economic relationships, such as marginal pricing, can be developed from the production 

function. 



260 

35 



In passing, it is amusing to note that perhaps the earliest paper in which 

marginal pricing is suggested is by one of the first energy analysts, Sir William Thomson 

20-21 
(Lord Kelvin). His solution to the problem of the most economical size of a 

14 
copper conductor for electrical transmission is still in use, and, as Smith pojnts ouf, 

this preceded the precise formulation of marginal productivity theory by Walras (in 

the fourth edition of his Elements ) by nineteen years and the formulation by V/icksteed 

(in An Essay on the Coordination of the Laws of Distribution) by thirteen years. 

One of the first serious attempts to analyze the connection between the economic 

production function and a process description more firmly grounded in engineering 

8 9 
practice was that of Chenery. ' However, there are several earlier and contempor- 
aneous papers in which technological information is specifically incorporated into 
economic treatments of production. 

Briefly summarized, Chenery conceives of production as the result of a series of 
industrial processes. Each of these processes involves a change, which is most usually 
a change in form, of input materials effected through the application of energy. The 
industrial plant is a map of material flows stimulated by applications of energy. In a 
parallel way, the first step of an assessment of the energy requirements of a complex 
process involves setting up a complete materials balance and then evaluating the 
energy flows that overlay this. Chenery's engineering production function is a long-run, 
ex ante, microeconomic production function. Thus, all input quantities are variable 
and functions of engineering design parameters. This means that trade-offs — substitutions- 



261 



36 



that could only be made by complete system redesign are considered. One of the 
most useful aspects of energy analysis is that it provides a means of evaluating just 
that sort of substitution process. 

Chenery sets up the following stripped-down definitional structure. Let 
X = f(u., . . . , u ) be the economic production function, in which X is the (optimized) 
output quantity and the u. 's are input quantities. This set spans all the goods of the 
economy. The amount of each input is determined by the set of equations 

u. = u.(v., ... /Vn ) 
in which the v.'s are the engineering variables, the physical parameters that ultimately 
specify the process. The economic production function combines this set of equations 
and the engineering production function, which has the form: 

X = * (v., ..., v n ) 
Chenery discusses determination of this function for the compression and pipeline flow 
processes of natural gas transmission. 

Following this work, Smith, Dan^, Manne, ' ' and Johansen have 
published more extensive treatments of industrial production models. Smilh's work 
has evoked substantial interest because he employs the long-run engineering production 
function as the foundation for a theory of investment. Smith is also apparently the first 

to consider productive systems that are best described by kinetic engineering production 

22 
functions. These functions describe the dependence of flow quantities of output 

(rates of production) on input stocks, and this mixture of stocks and flows in one equation 



262 



37 



is a strange bag for one trained in traditional economics. Very recently, Marsden 

18 
etal. investigated the forms of production functions arising from rate processes, 

and applied this apparatus to river water quality problems. 

Currently, the only empirical energy-analysis model for the total US economy 

23 
is that of Herendeen. This input-output model effectively incorporates a production 

function with fixed coefficients and no substitution, but only the energy portion of the 

production relationships are given. He has recently updated this matrix to reflect 

24 
the technological coefficients from the 1967 data, and this permitted a test of the 

projection procedures that had been employed. It is fair to say that the projection 
procedure that had been used, which consisted of deflating 1967 dollars to 1963 in 
the sector rows and renormalizing the energy rows by the energy/GNP ratio, did not 
appear successful. In this test, projections of the 1967 coefficients from 1963 
data were compared to empirical 1967 energy coefficients obtained by sectoral energy- 
to-dollar conversions. Apparently even over this four-year period of relatively smooth 
economic growth substitution is important, particularly in certain sectors. Another 
difficulty associated with use of the energy input-output matrix arises from the datedness 
of the raw input data, which takes a number of years to surface from the inner sanctums 
of the government. Also, its high degree of aggregation, even when the economy is 
broken down into 362 industrial sectors, makes it primarily useful for macro 
policy. Hopefully, we will soon find it possible to develop an input-output matrix 



263 



38 



in which the coefficients are obtained from process energy analyses, perhaps for 
both marginal and average technologies. 

In this and the prior section, we have attempted to sketch the manner in 
which physical information enters traditional economic production theory, with primary 
attention to energy. In the usual theory, the firm is considered to have knowledge of 
its short-run economic production function or the marginal products of each resource. 
Acting as an instantaneous profit-maximizer and price-taker, it adjusts its use of each 
resource to the level at which the resource cost is equal to the value of its marginal 
product. The data from energy analysis consists of the inverses of the marginal 
productivities of energy. These are incorporated into the economic analysis of the 
firm when the entrepreneur sets output levels at the point at which the resource unit 
cost divided by the output price is equal to the marginal productivity. The marginal 
productivity theory of short-run operating behavior must also be compatible with a 
description of the economic agent who, confronted with a long-run production function, 
acts to maximize long-run profits. This is the area in which physical information also 
has a natural entry into economic behavioral relationships — through the entrepreneur's 
consideration of the broad spectrum of production options implied in an engineering 
production function. 

Also, energy analysts have noted that one of the advantages of their methodology 
is that it permits an examination of tradeoffs between processes that form portions of a 
vertically integrated chain, but which are usually subjected to step-wise optimization. 



264 



39 



An energy analyst provided the following illustration. Admittedly, the example is 
oversimplified because it permits no adjustment to prices,and it *rs based on only a 
single resource. However, overall suboptimization resulting from the optimization 
of subsystems in real-world situations is well-recognized. Consider a .theoretical 
example involving a steel works and a car manufacturer. The operation of steel 
furnaces can be represented by material inputs of pig-iron and steel scrap, a fuel 
input and an output of steel, as shown in Figure 4. The pig-iron is produced in a 
blast-furnace which, let us assume, consumes E p units of energy per tonne of pig- 
iron. The steel scrap is not assigned an embodied energy, but the operation of the 
furnace requires E f units of fuel per tonne of steel throughput. Thus from ihe steel 
maker's point of view the total energy requirement per tonne of steel (E ) is given 



by 



E = E. + E (1 - &). 
o f p 



This suggests that the larger the scrap input, the larger $, then the smaller is 
the energy requirement of steel. In this hypothetical example, the steel manufacturer 
decides to install more scrap-handling furnaces and increases the price he is prepared 
to pay for steel scrap. 

At about the same time a car manufacturer is faced with a choice between two 
steel presses. Press A consumes 10 MJ per sheet pressed and rejects 10% of the plates 
as scrap. Press B requires 12 MJ per sheet, but doesn't reject any scrap. In the cause 
of energy conservation, and with the price of electricity rising (and ihe price of scrap 
rising) the price- and energy-conscious car manufacturer installs Press A. 



265 



40 







FUEL E./l 

1 


onne 




PIG IRON 


STEEL 
FURNACE 




_> ONE TONNI 


E / tonno 
P 






E / tonne 
o 



TONNES SCRAP 

Figure 4: The material and fuel inputs to a steel furnace. Note that to produce 

one tonne of output, the inputs are 3 tonnes of scrap and (1-p) tonnes of 



68-391 O - 76 - 19 



266 



41 



The net result of these two investment decisions is to increase the energy 

required to produce automobiles. This is illustrated in Figure 5 , which shows the 

fuel input to the press as E per tonne throughput and all the car manufacturer's scrap 

being used by the steel plant. If the output car requires 1 tonne of steel then, by 

conservation of mass and without slippage, the pig-iron input must be 1 tonne. 

However the fuel consumed in both the steel furnace and car press is proportional to 

the total mass throughput. This includes p tonnes of scrap in addition to the 1 tonne 

flowing from input to output. Thus the total energy requirement of the car, E , is 

E = E n + (1 + 3) (E. + E m ) . 
c p t m 

This shows that increasing the quantity of scrap generated and used, increasing (3, 
increases the energy requirement. If significant externalities of any kind exist, then 
it is well-known that step-wise optimization does not yield the most efficient solution 
for the total system. Again, information coming from energy analysis that could be 
summarized in an engineering production function could be utilized by a firm in making 
investment decisions that involve ^rtore than one process step. 



267 



42 



E f / tonne 



E / tonne 




Figure 5: 



1 TONNE 
AUTOMOBILE 
E 



he enlarged system including a car manufacturer who generates the 
eel scrap consumed by the furnace. 



268 



43 



OPTIMIZATION OVER TIME 

The subject of the optimal rate of use of resources has prompted intense and 

25 

sophisticated interest in the economic community. The history of this topic can be 

o/ 27—^5 

traced back to an early paper by Hotelling, through a number of published papers. 

The formulation that is generally employed is one drawn from optimal control theory 

(vide infra), in which an exogenously specified social welfare functional is maximized 

over time. A number of models assume that this functional takes the form: 

V = / e" pt u(c)dt 

r 

in which p is the discount rate on utility and u is the utility of consuming c at time t. 

35 
The treatment of Dasgupta and Heal clearly points out that the more general problem 

is one of joint optimization in a macroeconomic environment. How can one construct 

a program that provides efficient paths for both resource depletion and investment — 

investment that utilizes these resources? Of course, concern with the best societal 

use of natural resources moved a number of energy analysts to enter the field. Addressing 

this problem may be where energy analysis can most effectively enter economic description. 

Analysis of resource depletion must, at bottom, rest on the empirical analysis of physical 

production relationships, which can then be translated into value terms. In a sense this 

is a normative application of physical analysis. 

Faced with the possibility of total depletion of specific mineral resources, 

physical analyses are needed to define what substitutabilities between resources may be 



269 



44 



feasible, and to evaluate upper and lower bounds on the magnitudes of these effects. 
This information can be used by economists in determining not only efficient but also 

fair allocation of these resources over time. Economists are devoting increasing 

36-39 
attention to the normative question of intergenerational equity. 

The traditional approach to intergenerational equity issues has been to discount 
future utilities at some positive rate of interest, as indicated above. In the early stages 
of the Workshop, an energy analyst posed the following question for the economists. 
Consider nuclear generation of electricity and the problem of storing wastes from a 
nuclear reactor. This process must continue over tens of thousands of years, ond the 
storage process has energy requirements that may extend over the whole storage period. 
Also, take into account decreasing resource grades of energy resources, which result 
in a greater energy expenditure to produce an equivalent amount of fuel over time. 
In other words, | e t us examine the possibility that technological knowledge does 
not progress fast enough to overcome the rate of decrease in resource grade. In 
such a case, how should we discount the utility of consuming electricity generated 
by nuclear facilities over time? 

The economists' reply was that there are certain hypothetical cases, involving 
an essential but exhaustible resource and no technical progress in its utilization, in 
which discounting has objectionable consequences. In a model for the optimal depletion 
of an essential exhaustible resource, discounting future utilities favors an earlier 
generation over any surviving later generation. This effect increases monotonically 



270 



45 



25 
with the discount rate. This statement might apply to energy resources if renewable 

energy sources, such as solar or wind power, were not feasible. Another situation in which 
one might want to suspend discounting is when there is irreversible damage to an eco- 
logical system, such as the depletion of a species to a population below survival level or 
when harm is done to the health of future generations of mankind. In these situations, we 
are putting the responsibility on later generations for our detrimental actions, and 
this takes us beyond economics per se into the realm of human rights. In most other 
cases, so long as capital is scarce, efficient allocation of resources can be furthered 
by the use of a uniform interest rate. 

Energy analyses have recently been used in assessing the technological conse- 
quences of nuclear power construction scenarios. These were discussed informally as 
well as the energy analyses of the nuclear fuel cycle on which they are based. The 
values reported for the energy requirements for this process by independent investigators 
vary widely. The consensus was that conclusions should not be drawn from the evaluations 
available at the time of the Workshop because of this uncertainty. 

It was suggested that an alternative to carrying out such evaluations for ad hoc 

scenarios would be to formulate the assessment problem in the language of optimal 

40-41 
control theory. This, like any mathematical apparatus, forces us to state our 

question precisely, which often goes a long way toward determining the answer we get. 

In energy analysis, economics and thermodynamics, the first step is the deter- 
mination of what constitutes the system boundary. The relevant time-dependent behavior 



271 



46 



of the system is assumed to be completely described by specifying a finite number 

of variables called state variables. Each state variable may be a function of time. 

Let us abbreviate the set of state variables y. (t), . . ., y (t) in vector notation as 

y (t). One assumes that the time behavior of the state variables is controlled by a 

set of control variables (v,,.... v ) = v through a set of first order differential 
I m 

equations: 

4l ■ r <v, i »• 

The system's trajectory over time is then viewed as specified by the state variables, 
y. These are in turn determined by the control variables, v, which are not arbitrary 
but take on values over time, which are subject to deterministic control. 

Specifying the optimal control problem involves choosing some measure of the 
effectiveness of the control process, a function of v, y and t whose behavior over 
time is to be optimized. Calling this function F(v (t), y (t), t), and 

J tv, ?] = / F (v (t), y (t), t) dt 

t 

o 

we require that the integral J, called the objective functional, be optimized over the 

set of control variables, with y (t) evolving according to the set of differential equations 

above. The integral on p.48 is a special case of J. The specification is completed 

by giving an initial condition y (t ) = y . Constraints may be placed on the state 

o o 



variables independent of the control variables. 



272 



47 



This procedure has two advantages. The first is that precise specification 
aids in assessing the total implications of a particular option. Second, when the 
information is available/ we should choose among alternatives that represent efficient 
rather than non-efficient programs. Each efficient program is the optimized trajectory 
associated with a particular specification of objective functional and constraints. 

Optimal control problems can be formulated using only physical or economic 
variables. We can also use this apparatus to show one way in which the information 
from energy analysis can be introduced into economic optimization through defining 
constraints. For example, heat released by an electrical generation facility could 
have detrimental effects on a local climate. Because of this, the rate of heat released 
per unit area from the facility may be required to be less than or equal to a maximum 
acceptable amount. Physical analysis is required to define this constraint. A utility 
will choose to maximize its return (the objective functional) as a function of the rate 
at which it burns fuel in the face of this constraint. 

Some participants felt that another potentially fruitful area of interaction 
between technological analysis and economics lies in the theory of investment. The 
principal problem facing a producer in a market economy is the formulation of a plan 
for investment in capital goods that will maximize his long-run profit. The economic 
literature on investment is rich, this topic is one of a good deal of current interest, 
and it was not discussed in detail at the Workshop. 



273 



48 



14 
Smith grounds his monograph on the problem of investment in a technological 

description that utilizes an engineering production function motif, and this can be 

recommended as a source that gives serious attention to the question of the incorporation 

of physical information into economic theory. We are interested in properly describing 

the real-life behavior of entrepreneurs making investment decisions in a state of 

knowledge characterized by manifest uncertainty, particularly with regard to the 

introduction of technological change, the non-existence of markets for future delivery 

and market imperfections. Theoretical formulations of the investor's decision process 

that directly incorporate information derived from technological analyses should be of 

importance. 



274 



49 



SETTING THE LIMITS 

There was a strong consensus among both economists and energy analysts that 
one of the most promising applications of energy analysis and other technological 
assessments is in establishing the physical and ecological limits on economic processes. 
Such an effort would be useful in constructing intermediate- and long-range predictive 
models of large-scale systems. This information can be incorporated into econometric, 
input-output or systems dynamics models. In this application of energy analysis it may 
turn out to be unnecessary to augment the physical assessment with a valuation procedure. 
Energy analysis and other technological assessment procedures can be used without 
parallel economic analysis in testing the technological feasibility of societal options, and 
in rejecting non-feasible or unstable options. In a second stage, after the techno- 
logical feasibility has been ascertained, other inputs must be measured and included 
in any affirmative test of stability on an economic basis. Finally, the desirability or 
efficiency of the options can be determined using economic analysis. 

There are two different boundaries that can be established using energy analysis 
and technological assessment procedures, more generally. The first type of limit, 
familiar to most individuals who are concerned with predicting societal trajectories, 
is one which sets a maximum on a certain activity. Examples are easy to come by — 
the maximum pollution that can be absorbed by a lake before the limits on biological 
oxygen demand are reached, or the maximum amount of energy use that can be permitted 



275 



50 



before significant heating of the earth's atmosphere is observed. Maintaining the 
stability of the biophysical system implies certain limits on the maximum amounts 
(or rates) of use of resources. 

More surprising to the economist was the statement that a kind of lower limit 
can also be physically defined. All transformations — chemical, electrical or purely 
mechanical — will require at a minimum the energy needed for the reversible process, 
which can be calculated. Again, for a short discussion of the concept of reversibility, 
see Appendix II. For example, the minimum energy that would be required to synthesize 
a pelroleum hydrocarbon from carbon dioxide and water in atmospheric abundance in 
the year 2025 can be exactly assessed (and will be equal to ihe free energy of combustion 
of the hydrocarbon). This sort of assessment has strong predictive implications, and 
robust economic predictions must incorporate this kind of information. The delineation 
of upper and lower bounds is a tough problem, but it can be done in principle. The 
actual energy required for the process will be greater than the minimum because energy 
is needed to drive the process at a finite rate. 

To illustrate the possibility of comparing a real system to a theoretical limit, 
let us refer to Figure 6. This illustrates the gains in energy efficiency over time in 
the production of ammonia from methane and air. As can be observed, this process is now 
operating moderately close to its thermodynamic limit, which is quite exceptional for 
industrial processes. Note also that it is possible to leap beyond any particular thermo- 
dynamic limit by going to a completely new technology for producirg ammonia that 



276 



51 



100 _ 



Figure 6: Gross energy requirements (GER, see Appendix I) for the production of 
ammonia. Base reaction: 



100 



00 



30 «, 



CH, + AIR = NH- + C0 
4 3 2 



MJ 

Thermodynamic limiting GER - 17.5 r — 



THERMODYNAMIC LIMIT 



1910 



T" 

1920 



T" 

1930 



1940 1950 



1960 



1970 1980 

*NEW PROCESS 



277 



52 



does not employ reaction of air with methane. The thermodynamic limit for the new 
process can then also be calculated. However, it should be apparent that there are 
only a limited number of processes that utilize materials as abundant as air and natural 
gas (formerly), so that a feasible set of technology modifications ( e.g. , production of 
NhL from the direct, catalyzed reaction of nitrogen and hydrogen) could be projected, 
and the minimum energy for each transformation could be evaluated. But no matter how 
we refine our technology, there is always some absolute thermodynamic minimum energy 
requirement for the production of any good, even if we allow ourselves free choice among 
all available inputs. 

Let us explore a stability analysis in terms of physical variables. "Stability" 
here means that the flows and stocks represented by the variables remain bounded during 
the entire period of interest, which can be extended at will. With judicious caution, 
we can temper rigor with realism to use the theory of idealized nonlinear behavior to 
study real problems, provided we do not inadvertently push our models to times when 
our hypotheses are no longer valid. V/e cannot push our models to infinite t ime, for example, 
Like all stability analyses, these analyses give us negative information only, in the 
sense of defining ranges of parameters, such as intensitivities of use, within which our 
life styles must remain. The physical analyses will only be useful if they sometimes 
set tighter bounds on the regions of stability than do the price-based bounds. Whether 
this will be the case is not yet known. 



278 



53 



One particularly simple example is that of a physical resource that is slowly 
being consumed — vacant, primitive land, for example. The shadow (or real) price 
of this resource is initially zero or even negative, but ils price will be positive when 
some threshold is passed. This is also the cose for air and water as waste receptacles. 
These are the classic situations giving rise to externalities. Historically, it appears 
to be very difficult to stimulate or justify the scientific and technological work 
necessary to meet forthcoming nonzero prices of presently unpriced resources with enough 
lead time to meet the new scarcity in a technologically effective way. This is a state- 
ment in the language of classical economics of the simple, cynical truism that people 
won't start thinking about a problem until it already hurts to live with it. 

Stability analysis is a recognized tool for economic study. We are not aware 
of the application of this approach in the context of resource use with empirical 
production functions and resources constraints, although there have been model treat- 
ments of the allocation of a single resource, both exhaustible and inexhaustible. 
With the data from physical analyses, it will be possible to use technical production 
functions that reflect current or alternative engineering practices to decide whether 
a real or potential pattern of resource use is stable within the selected set of constraints. 

By way of illustration only, let us use a model for the system of production of 
an energy source (presumably coal) whose stock we denote by E, and one other resource that 
we may take to be steel, whose stock is S. These stocks represent inventories, not 
ultimate reserves. For example, E may represent already mined coal, or known 



279 



54 



reserves in mines, but not the total projected resource. For present purposes, take 

42 43 

as the model a system invented and studied in detail by the Brussels school (and 

44 
hence called the "Brusselator") and others, and extended to include a broader range 

45 
of possibilities than were encompassed in the original work. Our simple illustration, 

based on the Brusselator, is the production of steel, waste heat and waste matter from 
ore and energy reserves. 

Reserves ■ ) Energy Inventory (1 ) 

Ore + Energy ^ Steel + Waste Matter (2) 

2 Energy + Steel > 3 Energy (3) 

Energy > Waste Heat (4) 

The coefficients of energy in Step (3) need not be 2 and 3. These are simply the values 
for which the system has been most thoroughly studied. It is necessary that the system 
be nonlinear because of the real requirements for energy in the acquisition of energy. 
This nonlinearity lends both mathematical richness and portentous significance for policy- 
making to this problem. The equation* relating stocks to flows (not flows to flows or stocks 
1o stocks) take the form (when scaled to reduce the number of parameters to a minimum): 

^= r - mE + 6E 2 S - y E (5) 

jp= mE -6 E 2 S, (6) 

in which r = flow of energy reserves, and m = flow of ore reserves; 6 and y are parameters 
of the system, equal to the* rate coefficients for steps (3) and (4) respectively. Step (4) 
involves the final evolution of waste heat from energy. Equation (5) states that the rate 



280 



55 



at which the coal inventory used is equal to the rate of step (1), r, plus the rate 
6 Eb at which more energy resources are developed, less the rates of steelmaking / m E, 
and heating, y E, shown by steps (2) and (4). The convenient but not necessarily 
realistic assumption is made in Equation (3) that we develop three units of new coal 
inventory for every two consumed in development. Note that both 6 t S and y E are 
rates of flow and that the units of the rate coefficients 6 and y put all these quantities into 
common units. Note also that the stocks E and S in (5) and (6) are inventories of 
intermediate factors, whereas r and m are flows. The units of the rate coefficients 
are such that the steady-state condition on the inventories, dE/dt = dS/dt = is a 
necessary condition for economic equilibrium. The analysis of stability consists of asking 

what happens to the system if it is displaced slightly from the steady state. 

46 
The analysis of the Brusselator shows that the steady state may be unstable in 

47 
the Lyapounov sense if m/6, the flow of the reserve of ore relative to the energy 

2 
production rate coefficient 6, exceeds a critical rate of the form [ (r/m) + y/ 5 3. 

However, the situation for energy use need not be altogether bleak even if the 

Brusselator model were realistic and ore consumption exceeded its critical value, 

because at least one limit cycle always exists in the unstable region. This means that 

the energy-using technical production system may have a long grace period, during 

which the rates of energy consumption and steel production oscillate but remain bounded. 

During this time, the system can readjust its technology to move to a stable range or 

extend the life of its limit cycle. The price of operating with the rate of extraction of 



281 



56 



ore above its critical value is having to live with a cyclic, rather than a steady-state 
condition. Empirical energy analysis would be the means to evaluate the parameters 
m, r, 6 and v, and, at one level deeper, the coefficients in the relation (3). Let us 
once more emphasize that this particular simple system is meant to illustrate a direction 
for examining stability from the technical data of resource analyses, and has not been 
derived from observed data. 

The aim of energy analysis in this area must be the evaluation of realistic 
estimates within which our economic system must operate. We cannot be satisfied 
with the theoretical knowledge that limit cycles are a formal possibility. We must 
have a quantitative idea of what rates of resource use are stable or consistent with 
limit cycles, what rates are clearly inconsistent with either, and, within the limit 
cycle scenario, how long we could expect the cycle to endure. The "empirical 
economics" we call energy analysis (or resource analysis) is precisely the tool for 
this job. 



68-391 O - 76 - If 



282 



57 



THE ECONOMISTS' CRITIQUE OF ENERGY ANALYSIS 

On the final day of the Workshop, there was a period in which the economists 
voiced their perceptions of the strengths and weaknesses of energy analysis. They 
believe that the field could be strengthened in three major respects. 

A) The viewpoint of the discipline should be broadened to include other 
resources, and technological analysis in a larger sense. One group of 
energy analysts responded that a program concerned with water resources 
employing similar evaluative methodology had already been initiated in 
their group as a step toward a general multi-resources program. Water 
was chosen, they said, because they sense that it will be the next resource 
to suffer severe supply constraints. 

B) In a similar vein, the energy analyst should not be so concerned with a 
particular methodology. Rather than adopting a definite set of questions 
to which the discipline is pledged to seek answers, it should reorient 
itself to ask what questions are important in making societal decisions? 
The straightjacket of formalism should be shed. The energy analyst's reply 
was that he does not see himself as operating under such strictures and that 
the definitional structure suggested in the guidelines (Appendix I) was 
adopted primarily to ease communication in the field. Furthermore, the 

the methodology parallels that of thermodynamics, which permits evaluation 
of efficiencies and limits as discussed above. 



283 



58 



C) To be of maximum value / some valuation procedure should be adopted . 
It was suggested that the data be organized and spelled out so that an 
economist who works with an input-output or process model can incorp - 
orate it directly. The economists-Jielci- that energy analysis paled 
compared to economics as an allocator of scarce resources, and 
that this is not a proper function for energy analysis. The importance 
of consumer reaction and modeling consumer behavior — about which 
energy analysis hqs little say — was emphasized as well as the efficacy 
of introducing a price system, even if one doesn't believe in its optimality, 
in order to introduce behavior, particularly in making decisions over 
time periods. The rejoinder by the energy analyst is that pragmatically 
there are certain cases in which energy analysis may furnish faster or 
earlier signals than does economics, and that these situations should 
be empirically evaluated. He admits that in order to use energy analysis 
for any allocative function, some explicit or implicit valuation must be 
introduced. He also admits that any allocation procedure based on a 
single resource is inferior to that of efficient economic allocation. 
However, under conditions of energy supply constraint, governments 
may wish to evaluate the implications of policies on energy use. Further- 
more, based on this evaluation, considerations of national defense or 
similar goals may dominate the desirable goal of efficient allocation 



284 



59 






of resources. Several energy analysts are actively developing 
Input-output tables from process data. 









285 



60 



THE ENERGY ANALYSTS' CRITIQUE OF ECONOMICS 

One sense that some feei that the existence of the field of energy analysis 
is in itself a criticism of economics. Energy analysis is directed toward providing 
information for planning decisions, both those made outside the operation of the 
price system, and those where the market is not operating efficiently or with sufficient 
promptness. In this sense, the two systems of analysis are not antithetical, but 
energy analysis provides important inputs to economic analysis. For instance, energy 
analysis can provide a check on market operation by verifying the marginal pricing 
assumptions for this one resource. Also there are some economic occurrences that 
may be signaled more rapidly and v/ith equal accuracy by energy analyses. 

Thus, the principal comment directed at economics from the energy analyst's 
corner is that there should be a greater attention to the gathering of physical information 
appropriate to economic analyses and its incorporation. This is not to say that economists 
do not utilize such data or are unwilling to do so. As noted above, a number of economic 
studies have incorporated technological data directly, and the economists present emphasized 
the need to do so. Technological information is vital in developing accurate medium — 
and long-term models of economic systems, particularly in projecting technological change. 

As is the case with any healthy discipline, the sharpest attacks come from 

48,49 
within. There have been recent broad criticisms of the general equilibrium framework, 

by which is meant the rigorously-derived mathematical structure that is proposed as 



286 



61 



as a description of market operation under static or stationary-state conditions, as 
well as a response to these. The main thrust of this critique is behavioral, 
emphasizing the discrepancies between observable market processes and the body 
of assumptions and proposed behavioral mechanisms that are required in order to 
develop a closed equilibrium system. The energy analysts emphasize only the need 
to modify or go beyond the equilibrium model, rather than the rejection called for in 
the economic criticism. ' Some of the questions for economists from energy 
analysts are within the province of this behavioral challenge, while others probe only 
the necessity to extend the elegant equilibrium theory in describing the real dynamic 
system that may be often characterized by disequilibrium. 

1 ) Is the general equilibrium assumption of market agents who operate almost 
exclusively on the basis of price information accurate or should other direct 
Information bases, such as quantity, be incorporated? 

2) Does economic theory handle the problem of the dynamic evolution of social 
systems in an adequate manner? To what extent should exclusive control by mar 
forces be allowed in the face of physical constraints and the irreversibility 

of some decisions? 

3) Can we accept a description of the economic arena that goes no farther than 
an equilibrium analysis when we observe that it is usually in a disequilibrium 
state? As noted above, the emphasis in thermodynamics is on the transitions 
due to disequilibrium rather than en equilibrium states. 



287 



62 



At least one economist indicated that he was willing to support economics'current 
position with respect to each of these issues. 



288 



63 



ECONOMICS - ENERGY ANALYSIS INTERFACES 

On the positive side, there were a number of suggestions from both economists 
and energy analysts of areas in which greater interaction between energy analysis 
and economics is possible. Some of these have been implied in the foregoing text, 
and some are suggestions that are promising but were not considered at length. 

1) The use of technological analysis in designing predictive economic 
models, especially for long-range planning. 

2) The use of energy analysis in testing for viability of proposed systems 
of production. 

3) The development of sophisticated descriptive process models that will 
help an economist frame more realistic descriptions of technological 
change than a summary as an exponential function of time, a proxy 
for all time-dependent residuals. 

4) The determination of marginal input-output coefficients, for both marginal 
increases and decreases in output, In order to determine possible responses 
to sudden exogenous changes. 

5) The use of energy analysis in analyzing the operation of those economic 
sectors where there is government intervention and planning and also for 
better behavioral understanding of those sectors in which there is no 
direct intervention. 



289 



64 



6) The use of energy analysis in determining the relationship between the 
rate of a production process and its utilization of energy. By rate of 
production is meant the ratio of stocks of goods in process to the flow of 
output. One point that struck the economists as potentially important 
was the comment from energy analysts that less energy may be required 
per unit output as the rate of the productive process decreases or the 
duration of the process step increases. This indicates that under conditions 
of capital saturation, there is the possibility of decreasing the consumption 
of a resource by increasing the duration of the process step. 

7) The use of energy analysis in properly defining the characteristics utilized 
in a consumer production function. The utility of employing a char- 
acteristics space rather than a commodity space lies in the reduction 

in dimensionality, which is important in computer modeling. For example, 
in the heating of buildings the two primary characteristics would be 
heating convenience and heat energy. 

In summary, both the economists end the energy analysts felt that the data being 
generated by energy analysis and the dctciled process description can be of substantial 
value in sharpening the economic description of the system. 



290 



65 



ISSUES FOR FURTHER THOUGHT 

Near the close of the Workshop a subcommittee drafted a list of issues that 
they felt had either been raised by the discussion or remained from the original agenda 
presentations. Some of them represent modifications and refinements in the statements 
of the initial foci, and it may be useful to compare them with the list found on page 12. 
Many could usefully be employed as bases for further interfacial contact. The issues 
are not necessarily in order of priorities as seen by the Workshop, and the formulation 
and language is generally that of the subcommittee. 

1. At what point in the continuum of production activities, from primary extraction 
extraction to final demand, must a valuation system be selected or 

inferred? 

2. Is process methodology the most valuable single contribution of energy 
analysis to economic analysis? 

3. Are there consistent methods for allocating the energy inputs to the out- 
puts of joint production which satisfy both energy analytic and economic 
criteria? 

4. What are the relative advantages and disadvantages of energy analysis 
done by a process model in comparison v to that utilizing a macroeconomic 
energy input-output matrix? 



291 



68 



the Workshop, and it was noted that variable coefficients will be incorporated. It 
was suggested that these coefficients be both dynamic and price responsive. Technical 
advances are reflected in increased capital and labor productivity with time, but 
not in modified raw materials requirements. 

In order to relax the assumption of constant input coefficients the basic model 
can be supplemented with an energy matrix that gives a detailed description of the 
consumption of different kinds of energy in the production sectors in physical units. 
Process analysis information from the other sub-models can be used to change the energy 
coefficients. Thus the model facilitates simulation of various growth paths of the 
demand for energy. In the net energy demand sub-model a detailed technical treatment 
of ways of meeting a given net demand is undertaken by indicating the allocation 
pattern of energy resources to different sectors. For a given demand, alternative 
sources of supply that are economically efficient within the constraints of technology, 
environment, and reliability of supply are investigated in an energy supply sub-model . 
The supply model may generate heavy capital requirements for some solutions that may 
be inconsistent with the assumed or planned development of the economy. On the 
other hand, these solutions may have a positive effect on the economy through stim- 
ulating more efficient energy use. The MSG model thus serves to check the overall 
consistency of the assumptions of the various technical scenarios. Another important 
point that was raised at the Workshop was that ihe energy embodied in Danish imports 
and exports should be incorporated into this model if certain questions are 1o be answered. 



292 



69 



2) Bent Elbek. "Energy Analysis of a National Economy." 

Closely related to the investigation reported above is Elbek's attempt to use 
energy analysis methodology to evaluate all primary processes in the Danish economy, 
process by process, in order to formulate a total picture of energy use in that country. 
Sectors investigated in varying degrees of detail are agriculture (farming, gardening, 
forestry and fishing), industry (iron, metals, paper, chemicals, stone and clay, food, 
transport equipment, machinery, textiles, and wood), transportation (foreign shipping, 
domestic transport, air transport, private transport, communications), services (e.g., 
schools, hospitals, libraries, supermarkets), construction, and residential. Graphical 
displays of variations in energy, capital, and labor requirements of the aggregated 
sectors over the period 1950-1972 were also presented. This study is unfinished but 
impressive. In particular, it is one of the first to treat service and residential sectors 
in a direct manner rather than employing a "money-energy" conversion. 

3) Ingemar Stahl. "An Input-Output Evaluation of the Energy Requirements 
for 1000 MW Forsmark I Light Water Reactor. " 

This presentation outlined the merit-order strucJure of Swedish baseload 
electrical generation capacity, pointing out that closer matching of the price structure 
to marginal costs of generation could lead to a clearer perception of the cost of oil- 
fired generation facilities. A detailed (42-sector) input-output study of the energy 
requirements for constructing the Forsmark I reactor was described. This analysis, 



293 



70 



which utilizes 1971 prices, assesses reactor, construction, and turbine energy require- 
ments, as well as estimating energy use at site. Fuel and electrical energy require- 
ments are maintained as separate entries. The values reported in this careful evaluation 
were of reasonable magnitudes, but there was criticism of the exclusion of energy 
embodied in imported material components, which Stahl estimates would add 40-60% to 
the total. No initial core fuel requirements were included. Also, it was noted that 
the use of conversions based on aggregate industrial sector energy/l<rona rctios can 
lead to substantial error because of the specialized nature of many of the reactor com- 
ponents. 

In a separate session, there was a general discussion of the energy require- 
ments for the construction of nuclear facilities and the generation of nuclear power. 
It appears that these requirements are significant, but the energy analyses that have 
been published to date exhibit substantial variations. The disagreements seem to stem 
from many of them being back-of-the envelope exercises, and from the use of only 
money-energy conversions in several studies. Many participants felt that a careful 
disaggregated energy analysis based on physical inputs should be carried out before 
policy suggestions regarding nuclear facilities based on net-energy arguments are put 
forward. 



294 



71 



4) Lawrence Klein. "A Summary of Methods of Introducing Variable 
Coefficients in Input-Output Models. " 

A brief but detailed presentation of various methods of introducing dynamic 
and price-sensitive coefficients into input-output models to produce richer descriptive 
possibilities was given. The technical input-output module of an econometric system 
can be used flexibly. In particular, observed coefficients can be modified by engine- 
ering considerations of new processes. These can then be combined with statistical 
coefficients for the rest of the system. The three econometric approaches that were 
discussed, in order of increasing generality, involved the use of Cobb-Douglas, constant 
elasticity of substitution (CES) and translog functional descriptions. The CES function, 
employed in the Wharton model, was picked for closer examination, and the use of the 
model in assessing the effects of economic constraints or policy instruments was developed. 

5) J. M. Leathers. "A Description of the Use of Energy Analysis in Dow 
Chemical USA. " 

The history of the use of energy analysis as a materials management technique 
at Dow dates from 1965 and the submission of a design for a balanced energy plant 
for expansion of chlorine production facilities. The design was accepted, and energy 
analysis has expanded so that the Dow Accounting Department now maintains financial 
and energy accounts side-by-side for all of its 600 productive units. This appears to 
be a particularly effective management method for a company whose inputs and outputs 



72 



295 



NUMBER OF COUNTRIES 




Fi 9 ure 7: Distribution of energy use per capita 
per year (1971 ) for 178 countries. 



296 



73 



are energy-intensive, and its use at Dow has made an impression in the crucial area 
of profitability. The Dow system was described at length, including their treatment 
of wastes and by-products. Principal advantages appear to be the sensitivity of the 
energy requirements to modifications in the production system and the ease with which 
surveillance can be maintained on changes in these requirements. Also, in times of 
volatile price fluctualions companies may find it to their advantage to analyze their 
operations in physical rather than financial terms. A striking aspect of the Dow 
analytic system is that it is nearly identical to the one agreed upon in August, 1974 
at the First IFIAS Workshop on Energy Analysis, although neither group was aware of 
the other's methodology until March, 1975. 

6) J. P. Charpentier. "Distribution of Energy Consumption in the World 
(1971; 1 78 countries). " 

A plot of the distribution of the number of countries having a given energy use 
per capita as a function of that energy use is shown in the accompanying figure. A 
similar distribution is found for population, with the following differences: 72% instead 
of 75% (class III), 6% instead of 3% (class I), and 22% (identical) for class II. It 
again was pointed out that although this plot is interesting, it v/ould be valuable to 
have an accompanying graph in which the energy embodied in imports and exports 
is computed in the energy per capita figures. The reference used for the data 
in this plot is the U.N. Statistics Handbook. 



297 



74 



7) Willem van Gool. "An Informal Survey of the Use of Energy Analysis in 
The Netherlands. " 

The energy supply situation in The Netherlands, and energy research and develop- 
ment activities were succinctly described. Two major problems must be faced: short- 
term dependency upon imported oil for transport and the leveling off of indigenous 
natural gas production in ca. 1978. Investigations of future supply possibilities included 
thorough assessments of wind and solar energy. Energy analysis has been useful in 
examining the consumption options facing the Dutch over the long-term, and there are 
those who feel that it should be used in societal planning decisions in order to avoid 
severe constraints. For example, the energy supply options may be seen as facing the 
paired constraints . 

Energy Supply Sources 

fossil fuels, geothermal, nuclear energy 

fossil fuels 

nuclear energy 

wind, solar, biochemical methods 

any source on a large scale 

8) Peter Roberts. "Some Interesting Energy-Use Correlations for the British 
Economy. " 

Several intriguing empirical results were presented, including a plot showing 
the energy-use -per-household as a function of income in the UK for the following sectors: 



Physical Constraints 
thermal load 
C0 2 load 
radioactive waste 
land and sea surface 
local influence on climate 



68-391 O - 76 - 20 



298 



75 



services, fuel, transport, goods, food, housing, and alcohol and tobacco. A plot 
of the log of value-added-per-unit-mass-produced versus the log of energy-require- 
ments-per-unit mass over a wide spectrum of British industry was also introduced. This 
plot was linear and striking in its lack of scatter. 



299 



77 



APPENDIX I 
GUIDELINES FOR ENERGY ANALYSIS 

This memorandum is a summary of recommendations adopted at the First 
Workshop on Energy Analysis held in Guldsmedshyttan, Sweden, 26-30 August, 
1974, under the sponsorship of The International Federation of Institutes of Advanced 
Study (IFIAS). Twenty participants from ten countries took part; they were all 
engaged in studying some aspect of energy, and almost all have been active in 
analyzing how energy and related resources ere used. A full report of the Workshop, 
with examples, has been published by IFIAS in Stockholm. The goal was the production 
of a set of definitions, conventions and standards to be recommended for general use 
by those working with ihe analysis of energy. The motivaiion was the need felt by 
the organizers and the participants to facilitate accurate communication in this fast- 
growing field, and to do this in a way that would make the information useful to people 
outside the subject. 

The following summary presents only the final recommendations; the logic of 
choice and the considerations of alternatives ere discussed in the full report. 

Title and Subject 



The title "Energy Analysis" is recommended for the endeavor consisting of the 
study of the energy, free energy, availability or any other thermodynamic quantity 
sequestered in the provision of goods or services. "Sequestered" is employed in the 
sense of "set apart, " to indicate that energy may be tied up in the finished 



300 



78 



good or in the process materials, in addition to the energy used to do the work of 
the process. The title is intended to cover both the evaluation of energy and other 
thermodynamic quantities and the study of the implications of the results of the cal- 
culations. 

Quantities and Units 



The quantities recommended for use in the presentation of energy analysis 
data are: 

a) the internal energy E; 

b) the enthalpy H, equal to E plus the product of pressure P and volume V; 

c) the Gibbs free energy G, which is defined as the enthalpy less the product 
of temperature T and entropy S, whenever it is feasible to evaluate this 
quantity. 

Conventionally, the heating value that is recorded for a fuel is its heat of 
combustion at constant atmospheric pressure, which is an enthalpy, and energy analyses 
will customarily utilize enthalpies in process evaluations. Thus energy analyses will 
most often be enthalpy analyses. It is recommended that the gross heat of combustion 
be used; i.e., the enthalpy of combustion of a fuel consisting of carbon, hydrogen and ' 
oxygen should be based on products that are gaseous CCL qnd liquid H«0, with reactants 
and products at 273.1 5 °K. 

The Gibbs free energy change should be evaluated when feasible. There will 
doubtless be situations in which the availability A, which is equal to H-(T .S)+ 

(P .V), and the Helmholtz free energy F (equal to E + PV) will be useful also, 



301 



79 



but they are not recommended as normal forms of reporting data. Evaluations 
utilizing net heats of combustion, based on a combustion process that has as a final 
product water vapor (rather than liquid), may be valuable at times, but a statement 
explicitly noting their use rather gross values must be included. It may be desirable 
to evaluate the internal energy E when it differs significantly from the enthalpy. 
The unit of choice is the joule (J) and powers of ten thereof (megajoules, 
MJ=10 J, for example) in accord with the Systeme International conventions. Metric 
units of mass are recommended. Thus, energy per weight of product is conveniently 

expressed in megajoules per kilogram, (10 joules per 1000 grams), which is equivalent 

9 6 

to gigajoules per metric tonne (10 joules per 10 grams). 

Use of the following units is strongly discouraged by the Workshop in reporting 

energy analyses in technical media: British thermal units (Btu), kilowatt hours (kwh ), 

and all units based on material composition and therefore of variable value, such as 

metric tonnes of coal equivalent (tee), short tons of oil equivalent (toe) or barrels of 

oil (bbl). However their place in popularized presentations was recognized; it was 

strongly recommended that if these units are used, the data should also be given in 

standard metric units. 

The System and Levels of Analysis 

The system is that portion of the universe chosen for study. The system must be 
carefully defined by specifying its boundary. The system boundary separates those 



302 



80 



activities that are part of the question under analysis and are contained within the 
boundary from those activities that lie outside because they presumably have neg- 
ligible impact on the question. Energy analysis often begins with a focus on a product 
and the process stage by which it is fabricated from material inputs. The system boundary 
can be defined so as to contain this process stage and none other, and energy analysis 
would then calculate how much energy is required to carry out this single step. But 
the system could be defined to account for the energy used to prepare material inputs 
that are themselves fabricated in prior stages. Another choice of system boundary would 
include the final process stage and the processes that generate the inputs to the final 
stage. A further regression would have the boundary enclose all of these activities 
plus those that produce the inputs to the stages that yield the fabricated first-stage in- 
puts. This regression can be continued upstream until the system boundary encloses 
stages that employ only raw materials. Downstream, the boundary may or may not in- 
clude discard or recycling. 

Defining the system boundary also requires answering the question: what are 
the inputs to the system? Should one, for example, include the energy requirements 
for producing the capital equipment used in the stages of production? This problem 
and the Workshop's recommendations concerning systems and their boundaries can be 
described conveniently using Figure 8, which designates levels of regression. Level 
1 is the level of direct energy input to the final process stage. An evaluation at this 
level would include fuels and electric energy supplied to the process but none of the 



303 



81 



energy requirements for prior steps, such as the generation of electricity. The con- 
sensus of the Workshop is that data from Level 1, which is sometimes useful for engi- 
neering purposes, is generally not sufficiently informative for decisions in those areas 
in which energy analysis is particularly cogent. 

At Level 2, one includes the inputs to produce the materials used in the process 
and to provide the energy used at Level 1 . For those inputs that are themselves manu- 
factured commodities, there is a further regression. Level 3 takes into account the 
energy requirements of producing the capital equipment, as well as the first regression 
of the requirements for input materials considered at Level 2. Level 4 and higher levels 
continue the regression in the same way. 

Where, in practical terms, does one stop in this regression? Sometimes one 
is only interested in a particular level, and sometimes it is impossible to carry the 
analysis beyond a partial evaluation at Level 2. The Workshop recommends that when- 
ever possible, analyses be carried back to the level at which the contributions are 
comparable with the uncertainties in the contributions from preceding levels. Currently, 
this will often mean carrying out the evaluation through Levei 2, or, sometimes, 
through Level 3. Frequently an analysis through Level 2 will include 90 to 95% of the 
energy requirements calculated through Level 4, so that analyses terminating at Level 2 
will be useful representations. 

The increments to the total energy requirements, per unit of product, tend 
to decrease in magnitude as one goes to successively higher levels even though the 



304 



82 



number of inputs increases with level. Hence it is frequently appropriate to use more 
approximate and aggregated methods as one takes higher levels into account. (Note 
that the notion that the contributions diminish as the levels get higher is an approx- 
imation; there can easily be small contributions to the total coming from low levels, 
and occasionally, a large contribution from Level 3 or even Level 4.) 

The Workshop emphasized the importance of specifying the level and the system 
boundary, in order to permit comparisons among different calculations. 

Definitions of Measures 

The energy (or free energy, etc.) calculated for the system of interest is called 
the energy requirement (or free energy requirement, etc.). The term "energy require- 
ment" is recommended even when the figures refer to gross heats of combustion, i.e, 
enthalpies, so long as the meaning is clear. Use of the term "energy requirement" avoids 
the possible confusion that could arise from terms such as "energy cost, " which could 
be taken to mean the money costs of the fuels for the system. 

Four measures of the energy (or free energy, enthalpy, etc. ) require- 
ments are defined: 

a) The conventional thermochemical changes in enthalpy (AH) and Gibbs free 
energy (AG) (also, but probably less frequently, the change in the internal energy, 
(AE)and Helmholtz free energy, AF, of all the chemical and physical processes that 
occur within the system boundary. These are evaluated by enumerating all the chemical 



305 



83 



and physical transformations, such as combustion, chemical reduction of oxide ores, 
or evaporation, determining the enthalpy change or Gibbs free energy change for 
each, and summing the contributions from each reaction and transformation. 

b) The Process Energy Requirement (PER) is the sum of the fuel energy supplied 
to drive all the process stages within the system boundary, which may include the pro- 
duction of inputs beyond Level I. 

c) The Gross Energy Requirement (GER) is the Process Energy Requirement plus 
the gross heat of combustion of inputs that have alternative uses as fuels. Whenever the 
GER includes any energy sources other than fossil fuels, care must be taken to specify 
how the energy embodied in fuelstocks is defined. For example, the variety of avail- 
able technologies for energy production from fissionable materials allows a large range 
in the embodied energy one attributes to unit mass of material, so the value given to 
the GER will depend on the technology to which the definition refers. 

d) The Net Energy Requirement (NER) is the Gross Energy Requirement, less the 
gross heats of combustion of the products of the process.* This quantity reflects the net 
amount of energy required by a process if the products are finally used as fuel. If the 
NER is being evaluated for a real fuel, it is important to define whether the values taken 
into account are those of the actual heat derived from combustion (in which some energy 
may remain in uncombusted material), or the ideal (thermodynamic) heat of combustion. 

* Note that the heat of combustion of a hydrocarbon is defined to be a positive quantity. 



306 



84 



The Workshop recommends f hat whenever depletion of resource bases Is 
the concern, the GER's, NER's and PER's be evaluated as free energies of combustion 
and so identified. To obtain free energy requirements, corresponding energy (enthalpy) 
requirements must be evaluated, and it would be helpful to have the energy (enthalpy) 
requirement figures available as well. An explicit statement should be made, as to 
whether the energy requirements are computed as free energies or enthalpies of com- 
bustion. 

Evaluations using the conventional thermodynamic functions are well-approx- 
imated in many systems by the PER, which employs only heats of combustion. However, 
in a number of industrial processes, a significant fraction of the total energy employed 
is energy given off in chemical or physical transformations of the material being processed 
(exothermic ity of reactions), which is not included in the PER. The PER, GER and 
NER are measures of how much we draw on our stocks of fuels, while the thermodynamic 
energy and free energy requirements encompass total changes in usable energy. 

Comparisons between a hypothetical ideal process and a real process are made 
by defining the ideal process, end by evaluating absolute and relative measures of 
their differences: 

e) Free Energy Waste is the actual free energy requirement minus the ideal 
free energy requirement (AG . - AG. , . ). Similarly, the energy waste is the 

energy difference between real and ideal processes. These quantities can be calculated 
using total thermodynamic changes, GER's, NER's,or PER's, and must be so identified. 



307 



85 



The process upon which the ideal requirement is based must be defined explicitly. 

f) The Waste Factor w is the ratio of the Free Energy Waste to the actual 

AG . ,-iG., , 
r . 4 / _ actual ideal > , . ., . . 

free energy requirement (w - -r-fz ), ana is thus a relative measure 

ideal 

of the difference between real and ideal requirements. Again, the Waste Factor can 

be defined using energy changes or free energy changes, and may be calculated using 

total thermodynamic changes, GER's, NER's or PER's. The efficiency measures defined 

in (e) and (f) are arbitrary, and other workers may wish to develop additional parameters 

that are particularly suited to their problem. 

At least two meaningful per-unit-product values can be calculated for each 

of the energy requirement parameters of AG, AH, GER, NER, PER, Free Energy Waste 

(or Energy Waste) and Waste Factor. One is the average obtained by dividing total 

requirement by total output, and the other is the marginal value. The marginal value 

of an energy requirement is equal to the derivative of the energy requirement with 

respect to the amount of product, evaluated at the level of the last unit of product - 

the requirement for the last unit of output. Most data reported thus far have been averages, 

but both average and marginal requirements are being evaluated, and one must specify 

which values are presented. 

Partitioning 

If the product of interest is in joint production with others, there is an ambiguity 
as to how to allocate input requirements among the outputs. The allocation of energy 



308 



86 



Inputs when a process generates more than one good or service is called partitioning. 
The Workshop recommends that, whenever possible, energy requirements be partitioned 
according to a physical parameter. With several fuel products, for example, it would 
be natural to partition energy Inputs according to the energy embodied in the various 
outputs. It is also helpful to report the total unpartitioned requirements, so that people 
using the data can devise their own partitioning schemes. Obviously, for some policy 
applications, one might wish to partition according to product money values. 

Further Definitions 

Direct Energy is the gross enthalpy of combustion of fuels plus direct electrical 
energy used in a process or process stage, equivalent to the energy requirements of 
Level 1 . 

Delivered Energy (or Delivered Free Energy) is the output of an energy-analysis 
system delivered to a consumer. 

Energy Intensity (or Free Energy Intensity) is the energy requirement per unit money 
value or product, such as megajoules per dollar value. 

All these quantities can be calculated using AG's or AH's, GER's, NER's or 
PER's. 

Process Analysis is analysis based on the vertical flow of materials to yield a 
specific product or small set of products, such as a house, an automobile or a bushel 
of wheat. 



309 



87 



Input-Output Analysis , as in economics, treats the evaluation problem as 
one in which multiple inputs yield multiple outputs. The term has been applied largely 
to linear (matrix) relationships, but it need not be so restricted. 

Gra phic Presentation of Process Analysis 

Data from process analyses can be presented in flow charts. The Workshop 
recommends that this be done and a set of conventions for the chart representation was 
adopted. The symbols are shown in Figure 9, and should be displayed in the sequence 
indicated in the Figure. 

Rectangle: name of process stage; 

Triangle with vertex in flow direction: the PER for the process stage just named; 

Cart or bogey: the energy requirement for transport in this stage; 

Diamond: the energy requirement for capital; 

O val: the name and amount of product from this stage. 

These representations may be used for diagrams displaying AG's or £H's, GER's, 
NER's or PER's. To include the energy embodied in input materials that could be used 
as fuels, in order to represent the GER in a flow chart, use an Upper-half Semicircle 
for the enthalpy of combustion of any input that has value as a combustible fuel. To 
represent the NER, by including the fuel energy embodied in the products, use a Lower- 
half Semicircle for the enthalpy of combustion of any output that has value as a com- 
bustible fuel. 



310 



88 



If one wishes to show direct electric energy explicitly, the quantity of 
electric energy should be written into a box atop the triangle containing the total 
energy requirement for the process stage or for the particular input stream in the 
process stage. 

To compile the data into a flow chart, one fixes the unit of product and 
works upstream, filling in the amounts of materials and energy that ultimately go 
into the supplying of that unit of product. Figure.10 is an example of a flow chart 
for the production of one tonne of aluminum. 

As v/ith all the generalized energy requirements, the units that go into the 
triangles, diamonds and carts of the flow diagram may be energies, enthalpies or 
free energies. 

Simple addition of all the numbers in the triangles, carts and diamonds 
(but not the amounts of electricity in boxes on the triangles) gives the PER. Adding 
the values in upper -ha If semicircles to the PER gives the GER, and finally, subtracting 
the lower-half semicircles to the GER gives the NER. 



311 



89 



















1 


CT> 








1 






1 














1 2 


E 


a. 


3 




1. 












T s 


cu 


'3 




CO 




. t 


co 






• 6 


E 


cr 




c 




O 


<u 






s*» 


Q. 


a> 


O 







• *" ! 


c 
o 


o 
o 

CO 

»-> 
a> 






CP o 


3 
ST 


"0 


Q. 




•* 








e; </> 


0) 








"o 


• > < 

• o> < 


!i2 

cr 

c_> 

< 






Direct e 
Machine 


5 
o 


O 
O 

-O 

c: 



"S 
e 


CO 

3 

a. 

c 


a> 
c 
UJ 

























































1 


, 






CO 






1 

T e 


a> 








CO 


Jro 


*o 








a* .9- 


cr> 
o 











! "•» 


c 
o 

> !2 

'3 
cr 

;5 


i 

<L> 






Direct ener^ 
Capital equ 


lo 

c: 

o 











> 
• -J 








CP 
UJ 

























































































1 










i. 












w 










«/> 










</> 


y 










c 


! oj 




-o 




o> 
o 


>> 













. > 


o 


o 


o 


3 
O 

CO 


01 

C= CO 










"o 


55 


01 


QJ 


>» 


^-►1 
























^» 


, _i 






CT> 


CP 












o» 




3 


CO 


o 




o> * 














cr 


B 


fc 


a> 


.± "5 










o> 

c 




; < 




CO 


<u 


q r~ 












LU 




jj 


' 


















1 

in 

c 

s 


!*- 




CO 

co 
a> 
<-> 

o 






i 


! \ 


f 












c 


















• > 

4> 










CO 

co 

a> 

o 












* _J 














a> 






O 


















e 




1 O 








CL 










UJ 















312 



90 



Name of 
process stage 






Figure 9: Symbols for energy analysis flow diagrams. 



313 



91 



Bauxite mining 





Electrolytic 

reduction 

(Hall -HerouM Process) 




Figure 10: Typical flow diagram for production of 1 metric tcrtne of aluminum. 

9 
Energy units are gigajoules (10 joules). Empty diamonds signify 

capital requirements that have not been evaluated. Figures are 

based largely on British production. For this analysis, 

PER = 1.6 + 16.1 + 0.5 + 0.7 + 23.6 + 17.3 + 25.2 + 0.01 

0.05 + 216 + 0.6 - 302 GJ/tonne; 

GER = PER + 50.4 = 352 G j/ronne; 



314 



93 



APPENDIX II 
A THERMODYNAMICS PRIMER 

Thermodynamics is that branch of physical sciences that describes changes 
involving the transformation of heat and work into each other. It focuses on the driving 
forces responsible for the changes and the circumstances under which change will be 
predicted to occur, i.e., with disequilibrium. 

The language used by the thermodynamic ist is much the same as that adopted 
by the economist in analyzing a Lyapounov growth model. First, a physical system 
is defined by a precise description of its boundaries and by specifying the interaction 
of the system wilh the rest of the universe outside these boundaries. The macroscopic 
state of the system is determined by a set of measurable properties, the state variables. 
For a gas, these would be: pressure, P; temperature, T; volume, V; composition, C; 
and energy, E. 

For a change in which the system passes from one state to another, the difference 
in any of these properties b etween the two states is independent of the transition path 
and determined solely by the states. An equation of state expresses the mathematical 
relationship between all independent state variables. For example, the equation of 
state for a dilute gas containing non-interacting particles is PV = nkT, in which n is 
the total number of particles and k is a proportionality constant. 

A system is in an equilibrium state if it returns to this (original) state after a 
slight and temporary modification in external conditions. The properties of a system 
in an equilibrium state undergo no observable changes, even 'over an indefinite time, 
unless perturbed by external changes. 



315 



94 



Let us examine one example of a system in an equilibrium state, that formed 
by the chemical reaction of methane and air that is utilized in ammonia production: 

3CH 4 + 2N 2 + 30 2 * »4NH 3 + 3C0 2> 

When separated from each other, methane (CH .), nitrogen (NL) and oxygen (0,J are 
in equilibrium states. Upon bringing them together under the proper conditions, they 
will react to form ammonia (NHL) and carbon dioxide (CO«). The system's equilibrium 
state will contain a mixture of all five molecules. 

The description of how a system gets from its initial state to the final state is 
called the transition path . A particular transition mode that is very important in 
thermodynamic conceptualizations is the reversible path, which is a succession of 
equilibrium states. Anything that can happen in accordance with physical 
laws does, and this path defines the most efficient mode of transition available to the 
system. Because the change is thought of as occurring over a succession of infinitesimal 
states, the duration of the changes is infinite — the transition is carried out infinitely 
slowly. By reversible is meant that the path can be followed exactly backward to the 
initial state without any change of state variables. 

Thermodynamics relates the set of state variables to the process variables heat, O, 
and work, W, which are the instruments of change, through two basic laws. Both of 
these laws are empirical, but they are well-tested. 



316 



95 



A. First Law of Thermodynamics 

The change in the energy of the system is equal to the heat absorbed by 
the system less the work done by the system, A E = Q - W. The change 
In energy AE is a perfect differential and is independent of path, depending 
only on the initial and final states of the system. This law is an equivalent 
formulation of the law of conservation of energy. 

B. Second Law of Thermodynamics 

This law grapples with the idea of non-conservation of some physical variable. 
In order to formulate it, we define a new state variable, the entropy, S, by 
specifying the change in the entropy of the system between initial and final 

states 

c _ - dQreversible 
A 5> = J -y— 

in which dQ ... is the differential of heat along a reversible path for 

reversible 

the system, and the absolute temperature, T, can be physically interpreted 

to be the average energy per particle. The second law states that the magnitude 

of the right-hand side integral is always greater than or equal to its magnitude 

over the actual path followed by the system: 

. c . -DQreversible ^ r dQg ctual. 

A S = fj- > f — 

An alternate statement of the second law is that for any change in a system, 

there must be an increase in the entropy of the universe, which ?s the sum of 

the changes in the entropies of $i# system arid evervthing outside it, its 

surroundings. 



317 



96 



Beginning with these simple definitions, we can investigate what it means 
to set thermodynamic limits. For example, let us consider a change occurring in a 
closed system, with no heat exchanged with the surroundings so that dQ = every- 
where along the path. Then AS > for any real change or A S = for a change 
conceived of as being carried out along a reversible path. In examining thermodynamic 
limits, it is convenient to define a few additional quantities that are also state 
properties because they are composed of state variables. The enthalpy of the system, 
H, is defined to be equal to the energy of the system plus the system pressure times its 
volume, H = E + P V, and the enthalpy change is AH = AE - A (P V). 

This is a useful quantity for processes occurring at constant pressure, say open to the 
atmosphere, so that AH = AE + P AV. Now, let us ask what is the enthalpy of 
a process carried out reversibly at constant pressure: 

AH = AE +P AV 



sy A 

l First Law 



AH = Q -W + P AV 
sy 

W = work done by system against surrounding - (Force)x 

(distance) = (P ,. ) (volume). Only pressure - 

v surroundings 

volume work is considered. 



) 



A H=Q- P av + p AV 
su sy 



P = P for a change carried out reversibly (and only 
su sy 
for a reversible change). 



A H = Q ... 

reversible 



318 



97 



Thus, for a system in which constant pressure is maintained, the enthalpy change is 
equal to the heat absorbed by the system when the process is carried out in the 
thermodynamic limit of complete reversibility. The enthalpy change for combustion 
under a constant pressure of one atmosphere is the value utilized by the energy analyst 
in assigning an energy content to fuels. 

There are three other thermodynamic state quantities that arise in setting thermo- 
dynamic limits, each defined for a frequently observed situation: the Helmholtz free 
energy, F = E - TS; the Gibbs free energy, G = E + PV - TS; and availability, 

A £ E + P ,. V - T .. S . Under conditions of constant 

surroundings system surroundings system 

temperature and process reversibility, the change in the Helmholtz free energy is equal 

to the negative of the maximum amount of work that the system does on the surroundings, 

AF = - W . : 

maximum 



A F = A E - A(TS) 

J (T = const) 
AF=AE-TAS 

J (1st law) 
A F =Q - W - TAS 

J (definition of S) 

A F = Q - W -T/^ reversib,e 

(reversibility) 



A F = -W 

max 



y 



319 



98 



The work term. W , may be composed of electrical and chemical as well 
max ' r 

as mechanical work. If one examines a reversible process under conditions of constant 

temperature and pressure, the PV term in the Gibbs free energy cancels the mechanical 

work component of W to yield 4 G = W 

r max ' maximum 

non -mechanical 

For a spontaneous transition of a system that begins and ends in equilibrium with its 

(constant temperature and pressure) environment, the free energy change for the system 

must be negative, AG < 0. Availability is an especially interesting measure 

system ' ' 

in that it allows one to assess the maximum heat and work that can be exchanged between 
a system and the surroundings as a result of the previously constrained system's returning 
to equilibrium with its environment. 

Building on this foundation, it is possible to define several efficiency criteria. 
"First-law efficiency" is defined 1o be the net work done by a working cycle of a system, 
divided by the heat absorbed from a high temperature source by a low temperature sink. 

By mental construction of an ideal heat engine, one can show that the ideal first law 

T hiah - T low 
efficiency is equal to — = , an identification that actually requires use of 

'high 

the second law. "Second law efficiency" is defined to be equal to the utilized amount 

of availability, divided by the availability in the state of highest availability. For 

the purpose of energy analysis, it has been suggested that a related parameter, the 

AG , , - AG ... 
ii , r . ii i I r- i _ actual reversible c , 

energy waste factor be defined, w — -r= . For most real 



AG 



actual 



processes, this factor will be significantly greater than zero because energy is expended 
in effecting the change at a finite rate. 



320 



99 



BIBLIOGRAPHY 



I. B. M. Harmon, A nnals of the American Academy of Political and Social 
Science, 410, 139 (1973). 



2. H. T. Odum, Environment, Power and Society, Wiley - Interscience, 
New York (I97*JT 

3. H. T. Odum, Ambio , 2, 220 (1973). 

4. H. T. Odum, "Energy, Value and Money, " mimeo (first draft), 
University of Florida, Gainesville, Florida, (July, 1974). 

5. P. A. Samuelson, Quarterly Journal of Economics, 73, I (1959). 

6. See also, T. C. Koopmans, ed., Activity Analysis of Production and 
Allocati on, Wiley, New York, Chapters Vll-X (1951). 

7. For a precise discussion, see T. C. Koopmans; "Maximization and 
Substitution in Linear Models;" in Input-Output Relations (Proceedings 
of a Conference on Inter-Industrial Relations held at Driekergen, 
Holland; Netherlands Economic institute, ed.; H. E. Stenfert Kroesse 
N. V. -Leiden (1953). 

8. H. B. Chenery, Quarterly Journal of Economics, 63, 507 (1949). 



9. H. B. Chenery, in Studies in the Structure of the American Economy, 
W. W. Leontief et al., Oxford University Press, Oxford, England 
(1553). 

10. There are some pre-World War II studies by Ragnar Frisch, Ivar Jantzen 
and others in the Skandinavisk Tidskrift fo'r Teknisk Okonomi , a journal 
that is no longer published. See also C. Hildreth and S. Reiter, 
Chapter XI, and T. C. Koopmans and S. Reiter, Chapter XIV of 
reference 6. 



321 



100 



II. G. H. Borts, Econometrica, 1952, 71. 



12. A. S. Manne, Scheduling of Petroleum Refinery Operations, 
Harvard University Press, Cambridge, Massachusetts (1956). 

13. A. S. Manne and H. Markowitz, eds., Studies in Process 
Analysis, John Wiley, New York, New York (1961). 

14. V. L. Smith, Investment and Production , Harvard University Press, 
Cambridge, Massachusetts (1966). 

15. S. Dan/, Industrial Production Models , Springer - Verlag, 
New York (1966). 

16. A. S. Manne, Investments for Capacity Expansion, M. I.T. Press, 
Cambridge, Massachusetts (1967). 

17. L. Johansen, Production Functions, North Holland Publishing Co., 
Amsterdam (1972^ 

18. J. Marsden, D. Pingry and A. Whinston, Journal of Economic 
Theory, 9, 134 (1974). 



19. T. C. Koopmans, Three Essays on the State of Economic Science, 
McGraw-Hill, New York (1957). 

20. Sir William Thomson, Report of the British Association for the 
Advancement of Science, Transactions of Section A, p. 527 (1881). 

21. Another great thermodynamicist, G. N. Lewis, published a paper 
on political economics, "A Plan for Stabilizing Prices, " in a 
Journal edited by F. Y. Edgeworth and J. M. Keynes. See G. N. 
Lewis, Economic Journal, 35, 40 (1925). 



22. Smith's analysis of chemical kinetic processes is somewhat in error. 
In general, it is not possible to determine the functional dependence 
of the rate of a chemical reaction on the reactant concentrations 
from a knowledge of the stoichiometric information contained in the 
chemical equation, as claimed on pp. 50-51 of reference 14. 



322 



101 



23. R. A. Herendeen, "An Energy Input-Output Model for the 
United States (1963), User's Guide, " Center for Advanced 
Computation Technical Report No. 67, University of 
Illinois, Urbana, Illinois (1973). 

24. R. A. Herendeen and K. Shiu, "Comparison of Methods for 
Projecting Total Energy Coefficients, " Center for Advanced 
Computation Technical Report No. 147, University of 
Illinois, Urbana, Illinois (1975). 

25. For an introduction to optimal resource use over time, see 
T. C. Koopmans, Cowles Foundation Paper No. 396, Yale 
University, New Haven, Connecticut (1973); published in 
Economic Structure and Development , "Essays in Honor of 
Jan Tinbergen;" H. C. Bos, ed.; North Holland; New York 
(1973). 

26. H. Hotel ling, Journal of Political Economy , 39, 137, (1931). 

27. K. P. Anderson, Journal of Economic Theory, 5, (1972). 

28. N. Vousden, Journal of Economic Theory , 6, 126, (1973). 

29. W. D. Schulze, Journal of Environmental Economics and 
Management, I, 53 (1974). 

30. N. V. Long, Journal of Economic Theory , \0, 42 (1975). 

31. T. C. Koopmans, Quarterly Journal of Economics , 78, 355 (1964). 

32. A. Ingham and P. Simmons, Review of Economic Studies, 42, 
191 (1975). = 

33. V. L. Smith, "A Theory of Exhaustible Resources with Substitutions, " 
and "An Optimistic Theory of Exhaustible Resources, " California 
Institute of Technology, Pasadena, California (1973). 



323 



102 



34. G. M. Heal, "The Ideal Consumption Profile for an Inexhaustible 
Resource, " University of Sussex, Sussex, England (1974). 

35. P. S. Dasgupta and G. M. Heal, "The Optimal Depletion of 
Exhaustible Resources, " London School of Economics and 
University of Sussex, Cambridge and Sussex, England (December, 
1973). 

36. J. Rawls, A Theory of Justice, Harvard University Press, 
Cambridge, Massachusetts (1973). 

37. K. J. Arrow, "Rawl's Principle of Just Saving, " Economics 
Series Technical Report No. 106, Stanford University, Stanford, 
California (September, 1973). 

38. R. M. Solow, "Intergenerational Equity and Exhaustible Resources, 
to appear, Review of Economic Studies. 

39. W. A. Brock and J. A. Scheinkman, "On Just Savings Rules, " 
University of Chicago, Chicago, Illinois (March, 1975). 

40. G. Hadley, and M. C. Kemp, Variational Methods in Economics, 
North Holland/American Elsevier, New York (1971). 

41. L. S. Pontryagin, V. G. Boltyanskii, R. V. Gramkrelidze and 

E. F. Mischenko, The Mathematical Theory of Optimal Processes , 
Interscience, New York (1962). 

42. I. Prigogine and R. Lefever, Journal of Chemical Physics, 43, 
1695(1968). = 

43. G. Nicolis, Advances in Chemical Physics, 19, 209 (1971). 



44. J. Tyso'n, Journal of Chemical Physics, 58, 3949 (1973); J. J. Tyson 
and J. C. Light, Ibid, 59, 4164 (1973). — 



45. J. J. Tyson, Journal of Chemical Physics , 62, 1010 (1975), and 
many references therein. — 



324 



103 



46. R. Lefever and G. Nicolis, Journal of Theoretical Biology , 
30, 267(1971). 



47. A. A. Andronov, A. A. Vitt and S. E. Khaikin, Theory of 
Oscillators Pergamon Press, Oxford, England (1966)!! 

48. N. Kaldor, Economic Journal, 82, 1237(1972). 



49. J. Kornai, Anti-Equilibrium , North Hoi land/ American Elsevier, 
New York (1971). 

50. T. C. Koopmans, American Economic Review, 64, 325 (1974). 



51. K. J. Lancaster, Consumer Demand, A New Approach, Columbia 
University Press, New York (1971). 



325 

Thermodynamics and Energy Accountancy in Industrial Processes 

(C. Cozzi) ■ 

Abstract A prerequisite for any energy saving action in an indus- 
trial plant is an account of the process energy consumption. The aim 
is to point out steps in which the energy dissipation is great and there 
is a good possibility of saving energy. However, some problems still 
exist. For instance, how to compare thermal and electric energy units 
and how to measure the energy released or absorbed during a chemi- 
cal process are areas in which agreement is lacking. In this paper 
we show how these problems can be overcome. The way here pro- 
posed of performing energy accountancy consists of considering only 
the active part of each energy input or output in the process, i.e., 
only that part of the different types of energy able to do work. 
During the process active energy is converted into inert energy, i.e., 
energy that cannot do work and, hence, can no longer be considered 
an economic resource. 

AIM OF THE STUDY 

In the energy studies that are being conducted all over the world, one of the 
issues on which attention is frequently focused is energy accountancy. Some of 
the earliest papers on this subject (Berry and Makino, 1974; Chapman, 1974; 
Slesser, 1974), discussed some of the questions that arise at the interface between 
thermodynamics and energy accountancy. In particular, problems connected with 
the alternative use of thermal and electric energy units and the significance of 
the free-energy function in energy accountancy were considered. 

In this paper we discuss the relationships between thermodynamics and energy 
accountancy. We show that the energy conservation principle, which was assumed 
to support energy accountancy (Herendeen and Bullard, 1974), does not alone 
constitute an adequate foundation for it, since the principle of the irreversibility 
of practical processes also plays a fundamental role. Some inconsistencies that 
occur if one considers different kinds of energy as equivalent are pointed out, 
and a possible way to overcome them is presented. Our proposed way of perform- 
ing calculations is based on the view that energy is composed of an active part, 
which can produce work, and an inert part, which is unable to do work. 

GENERALITIES 

From an economic point of view, the object of energy accountancy is supposed 
to be limited and consumable. However, energy is neither consumable, as stated 
by the first principle of thermodynamics, nor limited, since the practically un- 
limited amount of heat at room temperature stored in the environment is freely 
available. This paradox warns us that the object of energy accountancy deserves 
careful definition. In fact, different types of energy do exist, and from experience 
we know that they have different physical characteristics and therefore different 
economic values. 

A usual classification of energies is based on the kind of transformation during 
which energy is exchanged between a system and its surroundings. Thus mechani- 
cal, electric, chemical, and thermal energies are exchanged when mechanical, 
electric, chemical, and thermal parameters of the system vary. For instance, when 
saturated steam at 10 bar and 180°C becomes water at 1 bar and 2")°C — a type of 
process that occurs frequently in industrial plants — the following energies are 
lost: thermal, because temperature, the thermal parameter, varies between 1S0°C 
and 2o°C; mechanical, because spatial coordinates, mechanical parameters, vary 
as a result of the shrinkage of the specific volume from 194 dm 3 /kg to 1 dm 3 kg; 
chemical, because the number of moles in the different phases, chemical parameters, 
vary during the condensation from steam to water. In an isolated system, i.e., a 
system that is prevented from exchanging energy (and matter) with the environ- 
ment, any transformation involves redistribution of the energy content among the 
different kinds. In perfectly insulated dwellings, for example, the heating unit 
transforms all the chemical energy of the fuel into heat of air and objects without 
losing energy to the environment. 



C. Cozzi is in Corporate Research and Strategic Planning at Montedison in Milan, Italy. 



326 

Sometimes systems cannot undergo any transformation without external inter- 
vention, in other words, without an energy injection. In such cases it can be said 
that the energy contained in the system is unable to do work, i.e., to promote 
transformations. Since human activity, and man himself, can exist only because 
transformation can be directed toward some definite end, only those kinds of 
energy capable of working deserve to be considered as economic resources. These 
kinds of energies, and only these, must constitute the object of our energy 
accountancy. 

ACTIVE ENERGY 

A deeper knowledge of energy able to work, hereafter referred to as active energy, 
can be obtained from the second principle of thermodynamics. In fact, since this 
principle is an assessment of the yield of conversion of heat into other kinds of 
energy, it lays the basis for a breakdown of each kind of energy into: 

1. active energy, which is able to promote transformations in systems since it 
can spontaneously and completely convert into other kinds of energy, and 

2. inert energy, which is totally useless since it is unable to change into other 
kinds of energy. Such inert energy can be thought of as heat at room temperature. 

Properties 

Only the outstanding properties of active energy are mentioned here, with 
detailed thermodynamic considerations discussed in a later section. Practical 
transformations, i.e., processes occurring in everyday life, take a finite amount 
of time and are therefore irreversible. Their irreversibility causes a conversion 
of active energy into inert energy. This implies, then, that active energy is a 
limited and consumable resource — an appropriate object for energy account- 
ancy. Another important property of active energy is that it is a function of 
state. Therefore, the active energy content of a system depends only on its actual 
state, identified by a set of parameters — mechanical, electrical, chemical, thermal, 
etc., and not on the basis of the process by which such a state was reached. 

Active Energy Consumption in Processes 

As a further consequence of active energy's being a function of state, the 
calculation of its consumption in a given process does not require a detailed 
knowledge of the process. In fact, the active energy consumption of a process, 
AA V , can be obtained from the following equation: 

AA P = A + AA 8 -Ai (1) 

where A { and A are the active energy inputs and outputs of the system and A A, 
is the variation of the active energy stored in the system. In the case of batch 
cyclic processes, or of continuous stationary processes, AA a — and Equation 1 
becomes : 

AA P = A -Ai (2) 

In practice, in order to calculate AA V , one must first choose the limits of the 
system, taking care to include all regions where some energy dissipation connected 
with the process occurs, even if such regions are outside the physical limits of the 
production plant. For example, if warm water is discharged from a plant into a 
river, the thermal energy dissipation occurring in the river must be included in 
the accountancy, even though the river is physically outside the plant. Moreover, 
when polluting effluents are discharged into the environment, the conscientious 
energy accountant should extend the limits of the system to all the dissipation 
regions in which energy will be spent, whether by government or individuals, 
either to reduce pollution or to survive in the presence of it. 

After defining the limits of the system, the energy accountant determines all 
energy flows that cross the limits during the process, and assesses their active 
energy content, A { and A , relative to some arbitrary reference state, equal for 
each flow. Finally, AA S is calculated as the difference between the initial and 
final active energy content of the system. Since the active energy function is* 

A=U-T S (3) 

where U is inner energy, T Q is room temperature, and S is entropy, it follows that 

AA S =AU-T Q AS (4) 



*This magnitude is different from the well-known Helmholtz free energy, F= U—TS, since it does not 
depend explicitly on temperature. It is similar but not equal to the availability Av= U—ToS+PoV. 



327 

Practical formulas for calculating active energy content of energy flows are 
listed below: 

1. Heat flow at constant volume and constant temperature T. Such conditions 
occur when the heat content of the reservoir is very large in relation to the amounts 
exchanged, or in continuous stationary processes. The active energy content of the 
heat quality Q is 

A T =Q^^° (5) 

where T is room temperature. When temperature varies between T\ and T 2 during 
the heat flow and volume remains constant, the active energy content of the 
exchanged heat is 

(6) 



A T=§1[ C — T -° dT= A(T 2 )- A(T,) 



where C is the constant volume heat capacity of the reservoir. Equations 5 and 6 
hold for temperatures both higher and lower than T . 

2. Chemical energy flow, i.e., the flow of chemical compounds. The active energy 
content is calculated by means of Equation 3, relative to some arbitrary reference 
state. 

3. Mechanical energy flow. The entire mechanical energy is active energy. An 
example of this is the energy input of a hydroelectric power plant. When expansion 
work is given up, or taken up, by some reservoir in physical contact with the sys- 
tem under adiabatic conditions, i.e., without concurrent heat exchange, the 
following equation holds: 

A M = - T 2 pdV = A(T i )-A(T l ) (7) 

where p, V, and T are pressure, volume, and temperature, respectively, of the 
reservoir and 1 and 2 are its initial and final states. 

4. Electric energy flow. All electric energy is active energy. 

As an example of the calculation of active energy consumption in processes, 
consider the following practical process. One kilogram of water is heated at atmos- 
pheric pressure (1.05 bar) from 25°C to 50°C by mixing with saturated steam at 
10 bar and 180°C. The amount of steam required is approximately 40 g. If we 
consider the system as composed of 1040 g of water, we can write the initial state as 



and the final state as 



1000 g liquid at 25°C, 1.05 bar 
40 g vapor at 180°C, 10 bar 

1040 g liquid at 50°C, 1.05 bar 



From Equations 1 and 4 and the usual tables of thermodynamic properties of 
saturated steam, the active energy variation in the system can easily be calculated 
as 

AA = -21kJ 

This means that during the mixing process the system loses some of its ability to 
do work. The free-energy variation in the same process is 

AG=AH-ATS= -7AU 

The system being isolated, the enthalpy does not vary : AH = 0. 

The energy contents of some typical energy inputs of industrial processes arc 
listed in Table 1. As a reference state, liquid water and gaseous carbon dioxide at 
25° C and 1 atm were assumed. From this table it appears that AG can be used 
instead of AA for a characterization of common fuels. In fact, the two magnitudes 
have almost the same value when they refer to a chemical reaction that occurs 
isothermally at room temperature. For nonisothermal processes, however, AG 
can be misleading since, disregarding contributions due to chemical transforma- 
tions, it decreases when heat content increases. The value of AG calculated for 
the previous example, AG= —7.4 kJ, then lacks all value for energy accountancy. 



328 

TABLE l.-ENERGY CONTENTS OF SOME TYPICAL ENERGY INPUTS 



+32.8 


+32.7 


+51.2 


+55.7 


+46.5 


+48.4 


-.104 


+2.61 


-.201 


+2.67 


-.299 


+2.70 



Inputs AA(kJ/g) AG(kJ/g) AH (kJ/g) 

Coal (100 percent C) at 25°C +32.8 

Natural gas (100 percent CH<) at 25°C +51. 5 

Fuel oil (100 percent n = C 8 Hi 8 ) at 25°C +46.6 

Saturated steam at 2.5 bar... +.457 

Saturated steam at 10 bar +.625 

Saturated steam at 25 bar +.745 

Active Energy Content of Goods 

One aim of energy accountancy is to find energy values for different goods, 
indicating the amount of energy resources consumed in their production. One 
possible way to calculate such quantities is to identify networks of processes 
involved in obtaining the products (Chapman, 1974). Then, for each network, the 
energy consumptions of the processes are summed. When a process gives more 
than one product, the total energy consumption must be distributed among the 
various products. The result of the summation, which expresses the consumption 
of active energy of the network, may be called the irreversibility content of the 
product. This content is always a negative quantity: it represents the amount of 
active energy that has been irreversibly converted into the useless form, heat at 
room temperature, during the production process. This useless energy is not 
stored in the product, and, of course, it is not recoverable. 

A completely different energy parameter that characterizes the product is its 
actual active energy content. A reference state for mechanical, electric, and 
chemical systems being defined, this quantity can be calculated as the amount of 
active energy that the product can give up in a reversible transformation leading 
it from its actual state to the reference state. For thermal energy the reference 
state is represented by T , room temperature, as implicitly stated above. Since 
active energy is a function of state, the amount of it that is stored in a product 
does not depend on the production process, but only on the actual state of the 
product. Moreover, it may be either greater or less than the active energy content 
of the goods entering into the process as reactants. Steel, for instance, contains 
more active energy than the correspondent amount of iron ore, but sulfuric acid 
contains less active energy than sulfur. Finally, the active energy content of a 
product can be recovered, bringing the product to the reference state. 

A practical link between irreversibility content and active energy content, which 
are two rather heterogeneous quantities, is given by Equation 1, or, in the more 
common case of batch cyclic or continuous stationary processes, by Equation 2:* 

A =A { +AA P (2a) 

Thus the active energy content of a product, A , is given by the sum of the 
active energy content of the input and the irreversibility losses of its production 
process. Another way to write Equation 2 is 

Ai=A +AA v (2b) 

That is, the active energy input to the process is equal to the active energy content 
of the product minus the irreversibility losses of its production process. Since these 
two latter quantities are rather different in meaning, one should avoid considering 
the right-hand side of Equation 2b as the total energy content of the product. 

The above considerations are limited to the ideal case of a production network 
that consists of only one process. In actual cases, where complex networks must 
be considered, an important problem to be solved is finding a definite criterion for 
distributing irreversibility losses among different products. 

THE CHOICE OF A REFERENCE STATE 

As mentioned above, a reference state, to which a zero active energy content 
has been conventionally attributed, is essential in determining the active energy 
content of a product. Moreover, even if this state is not strictly necessary to 
calculate the irreversibility losses of the processes, AA p , it is of practical value 



'Here we assume that the output of the process is a single product. 



329 

for carrying out calculations. For mechanical systems the reference state is con- 
nected with the choice of the coordinate frame, and for electric systems it is often 
denned as a state of infinite separation among charges. No simple choice exists 
for chemical systems. Of course, a general chemical reference state should be one 
of very low active energy content, so that goods may reach states of negative 
active energy content only rarely. Furthermore, and of even greater importance, 
it should represent a common active energy level for the different chemicals; i.e., 
reversible transitions between chemicals at the reference state should proceed 
without any exchange of active energy with the environment. Probably a simple 
reference state fulfilling these conditions does not exist. Thus we are compelled 
to be satisfied with different reference states for different cases. 

In practice, a chemical reference state could be one of minimum active energy, 
the system being allowed to react with practically unlimited and freely avail- 
able chemicals, such as water and air, and to reach infinite dilution in the en- 
vironment. This being a practical choice, one is free to introduce further simplifi- 
cations, taking into account particular situations. For instance, in determining the 
active energy content of a fuel, it is reasonable to neglect the dilution process of 
carbon dioxide in the atmosphere or in water (even if in the latter process a further 
5% of the heat released by burning coal might be recovered), since at present no 
one thinks of recovering the energy of such processes and the opposite work of 
concentration is carried out in photosynthesis at the expense of solar energy which, 
within certain limits is freely available. On the other hand, in considering the 
active energy content of sulfur, the dilution in water of its oxidation product, 
sulfuric acid, should be explicitly taken into account, since in many industrial 
processes the dilution of sulfuric acid followed by its reconcentration is an im- 
portant step from the energy balance standpoint (as, for example, in the sulfuric 
acid process for hydration of ethylene to ethanol). 

THERMODYNAMICS AND ACTIVE ENERGY 

The Thermodynamic Meaning of Active Energy 

Having defined active energy A as the energy able to do work, we can find a 
link between A and the usual thermodynamic functions as follows. According to 
the first principle of thermodynamics, 

dU=dQ + dW (8) 

where U is the internal energy of the system and Q and W are the heat and 
work, respectively, entering into the system. Note that in Equation 8 heat and 
work are considered as different magnitudes, even if they have the same di- 
mensions. In fact, according to the second principle of thermodynamics, heat 
can be transformed only partially into work. More precisely, the heat Q 2 of a hot 
reservoir at temperature T 2 can produce the work W when a cool reservoir at T u 
lower than T 2 , is available, in which the heat Q^=Q 2 —W can flow. The yield of 
transformation is 

W=Q 2 T *-^ (9) 

When the unlimited amount of heat at To of the environment is considered, it 
may be observed that no work can be obtained from it, since there is no cooler, 
unlimited, and freely available reservoir to receive part of the heat.* It is there- 
fore correct to define this thermal energy as useless and to call it inert energy. 

However, a complete separation of heat and work, as in Equation 8, is not con- 
venient from our point of view. Since our aim is to compute the total work avail- 
able in a transformation, we should extract the potential work stored in dQ 
and add it to dW, thus obtaining the differential of the total work. According to 



•This is in fact only a first approximation, because natural temperature gradients exist in the environment 
and in principle could be utilized to obtain work. 



330 

Equation 9, and considering the environment as a cool reservoir, the amount of 
work corresponding to dQ is 

dW Q = dQ^^ (10) 

In reversible transformations. 

dQ=TdS T „ (11) 

where S is entropy. From Equations 8, 10, and 11 we can then obtain 

dW tot = dW Q + dW=dU-TodS re y=d(U-T Q S T ey)=dA (12) 

Equation 12 means that the total amount of work exchanged in a reversible trans- 
formation is equal to the variation of the active energy of the system. Equations 
3 and 4 follow immediately from Equation 12. The active energy A must be a 
function of state, since U and T S are. 

The above-mentioned considerations hold for reversible transformations. In such 
cases, if the process is reversed, one can completely recover the resources used. 
However, real processes are always irreversible. In this case, beside the heat flowing 
from the environment, which constitutes the only entropy source in reversible 
processes, dS Tev , consideration should also be given to the entropy production with- 
in the system due to irreversible processes, such as attrition of heat exchange at a 
finite temperature gradient, diS irr ev. 

From a thermodynamic point of view, all these facts indicate that the entropy 
of the system increases to a greater extent than could be foreseen according to 
Equation 11. In particular, if we put 

dS=dS w +dSi,m (13) 

substitute Equation 13 into Equation 12, and integrate, we obtain 

AW= A U- T AS+ T AS irrev (14) 

Comparing Equation 14 with Equations 1 and 4, and since AW=A { — A we 
find that 

A^ p =T A5 irre v (15) 

Equation 15 means that the active energy lost in a process because of irreversi- 
bility can be considered as transformed into heat at room temperature. This equa - 
tion also indicates a possible way of calculating the active energy losses in the 
processes. 

Active Energy Content of Energy Flows 

The active energy content of heat Q, exchanged at constant temperature T and 
constant volume (i.e., without expansion work exhange), follows immediately as 
a consequence of Equation 9: 

A^Q 7 ^ (5) 

If during the exchange process the temperature varies from T x to T 2 , Equation 5 
holds only for an infinitesimal heat quantity dQ = CdT, where C is the constant 
volume heat capacity of the reservoir: 

rp rp 

dA T =C L -~^ dT 
and, integrating, 

^ r= JT c ^V"° dT (6) 

Recalling the definition of A, Equation 3, one can obtain from the above equation 

A T =A(T 2 )-A(T 1 ) (6a) 

It can easily be shown that Equations 5, 6, and 6a hold for temperatures both 
higher and lower than T . In the latter case, considering for example Equation 5, 



331 

it can be observed that the flows of active energy and heat are opposite. In fact, 
as a refrigerator gets warmer it acquires thermal energy but obviously loses the 
ability to do work. This is a further example of the unsuitability of thermal energy 
as a basis for energy accountancy. 

The active energy content of nonthermal kinds of energy is equal to the total 
energy content. In fact, in principle nonthermal energies can be transformed 
completely into one another. In particular, for expansion work in adiabatic 
reversible conditions, A S is equal to zero and Equation 4 becomes 

&A = AU = -f*pdV=A(Vi)-A(V l ) = A(T i )-A(T l ) (7) 

All values of Equation 7 refer to the reservoir exchanging expansion work with 
the system. In nonadiabatic conditions the reservoir also would exchange heat 
with the surroundings. The active energy A would no longer be a correct represen- 
tation of the mechanical work exchanged, and the equation stating the complete 
interchangeability of mechanical energy with active energy, 

&A = MJ 
would no longer hold. 

According to the definition of A, the active energy content of a chemical system 
is given by Equation 3, which relates to some arbitrary reference state. A brief 
examination of the different energy functions that are normally used to characterize 
chemical reactions might be of some interest. 
Enthalpy, 

ff-L T -PF (16) 

is largely used to characterize reactions of combustion and therefore fuel. In 
fact, it represents the heat exchanged when the chemical reaction is carried out 
without exchange of work with the environment (except expansion work) and 
reactants and products have the same temperature and pressure. However, it 
does not lend itself to use as an active energy measure, since in enthalpy heat and 
work are mixed. Moreover, even if considered as a pure thermal magnitude, it is 
of limited use, since the temperature at which the heat is set free during the 
combustion is usually not specified. 



Free energy, 



G = H-TS (17) 



is mainly used, in the energy field, to characterize energy exchanged in elec- 
trochemical processes. It was originally introduced to describe equilibrium 
conditions of chemical systems at constant temperature and pressure. Fur- 
thermore, it represents the work available, excluding expansion work, during the 
chemical processes that occur at constant temperature. Adding to G, expansion 
work, and work content of thermal energy TS, one obtains Equation 3: 

A=U-T S (3) 

Free energy was proposed for use in energy accountancy (Berry and Makino, 
1974), and in fact it can be used as a first approximation to describe active energy 
variation in chemical reactions occurring at constant temperature. It does not 
lend itself to describing active energy variations during heating or cooling, since 
in this case G, disregarding contributions due to chemical transformations, 
decreases when the heat stored in the system increases (see Table 1). This depends 
on the fact that G has a negative temperature derivative: 

which warns that the work interpretation of AG must be carefully restricted to 
isothermal transformations. 



332 

The function "waste factor" (Slesser, 1974) 

A Gael - AGid 



WF = 



AG a 



is of little use, since it undergoes the limitations of G, and, moreover, in the case 
of AG ac t = (occurring for instance in spontaneous processes in which no free 
energy is either consumed or recovered) WF goes to infinity without corresponding 
to infinite energy consumption. 

CONCLUSIONS 

The energy accountancy method presented in this paper, when applied to all the 
single unit operations constituting a process, allows one to obtain precise indica- 
tions of the energy losses involved in the process and, moreover, enables one to 
point out critical areas where irreversibility losses are localized and possibilities of 
improvement exist. 

Since energy planning is a worldwide problem, the possibility of enlarging the 
system to an entire country or to the world should be considered, keeping in mind 
that what is gained in generality is lost in precision. In this connection, the mean 
values of energy consumption of unit operations of the different industrial branches 
could be determined. Such values would allow one easily to obtain a realistic 
assessment of energy consumption for industrial processes from a simple knowledge 
of the flowsheet. The obtainment of such mean values would simultaneously 
require an accurate analysis of the unit operations, according to the principles 
presented here, and the examination of available statistical data on energy and 
goods inputs and outputs of processes. 

REFERENCES 

Berry, R. S., and H. Makino. 1974. Energy thrift in packaging and marketing. 
Technol. Rev. 33-44. 

Chapman, P. F. 1974. Energy costs: a review of methods. Energy Policy. 

Herendeen, R. A., and C. W. Bullard III. 1974. Energy costs of goods and 
services. Paper presented at the IFIAS (International Federation of Institutes for 
Advanced Studies) Workshop on "Energy Accountancy" held in Guldsmedshytton 
(Sweden) on August 25-31, 1974. The proceedings were not published. 

Slesser, M. 1974. Energy accounting. Preliminary survey of the state of the art. 
Paper presented at the IFIAS (International Federation of Institutes for Advanced 
Studies) Workshop on "Energy Accountancy" held in Guldsmedshytton (Sweden) 
on August 25-31, 1974. The proceedings were not published. 



APPENDIX IV 
The Nuclear Power Debate 

Interest groups opposed to nuclear power in principle have sought 
to use energy analysis to demonstrate that the nuclear option is an 
energy drain rather than an energy suppl} T system. The debate in 
print unfolds in the articles which follow. The most recent publications 
show the debate narrowing to questions of optimum rate of construc- 
tion of nuclear power plants and grade of uranium ore worth exploiting. 



335 



ENERGY INPUTS AND OUTPUTS FOR 
NUCLEAR POWER STATIONS 



by 



P.F. Chapman 

and 
N.D. Mortimer 



Energy Research Group, 
Open University, 
Milton Keynes. 



RESEARCH REPORT ERG 005 > SEPTEMBER 1974 

(Revised December 1974) 



336 



Contents 



Preface 



Introduction 



2. Criteria for new sources of energy 

2.1 Energy profitability 

2.2 Transient problems 

2.3 Long-term problems 

3. Energy analysis of nuclear reactors 

3.1 Types of inputs 

3.2 Energy values of inputs 

3 . 3 The net outputs 

4. Power analysis of reactor programmes 

4.1 Principles of analysis 

4.2 Results for various building schedules 

4. 3 Future plans 

5. Conclusions . 

6. Answers to some difficult questions 
References 



337 



Preface 

This is an interim report of a research programme which has 
only been underway for 6 months. The aim of the report is 
to describe a method of analysing nuclear power systems so 
that some of the physical consequences of decisions can be 
understood. At the present time we are happy with our 
methods and, thanks to the co-operation of many others, 
reasonably confident of our data. However there are still 
large uncertainties in the data and the results are probably 
not accurate enough to allow detailed comparison of different 
reactor designs. 

There were a lot of reactions to the previous version of this 
report, some very helpful, others downright silly. The last 
section of the report sets out our answers to some of the 
questions which seem to recur when people first read this 
report. 

To clear away a silly point at the outset it is necessary to 
state that the authors of this report <?o not think that the 
evidence we present means that nuclear power stations should 
not be built. What we have attempted to evaluate are the 
energy implications of building reactors, that is all. We 
believe that this information should be considered by 
governments and institutions who formulate nuclear policy - 
but we do not think energy analysis should dominate policy 
decisions. 

We are extremely grateful to many individuals within the 
nuclear business for pointing our errors in our initial report 
and with supplying us with much better data. However, the 
responsibility for any mistakes or errors in the report is 
ours, we only hope that interaction with these people has 
made the results section more realistic. We would particularly 
like to thank L. Brookes (UKAE) and R.D.Vaughan (TNPG) and his 
colleagues who have both provided important information. We 
would also like to thank J. Price (F.o.E) for his careful 
analysis of the paper and some important extensions of the 
problem of energy investment. 



338 



1 . Introduction 

Since the first demonstration of controlled nuclear fission 
reactions the energy available in fissile material has been 
viewed as a medium and long-term substitute for fossil fuels. 
The interruption of oil supplies in 1956 gave a considerable 
boost to the UK nuclear programme. There has been a growing 
pressure to give another boost, in the form of planning 
permission and funds, since the interruption of oil and coal 
supplies in 1973. Since there is a significant probability 
of further interruptions and price rises for oil and coal 
the pressure to invest in nuclear power is likely to 
increase over the next few years. Thus although the U.K 
government decision of July 1974 resulted in a much slower 
development, this decision will continue to be threatened 
and may well be upsurped. 

Most authorities that have looked at the long term prospects 
for the U.K, and for most industrialised countries, have 
stressed the importance of nuclear power. Every projection 
of future energy demand shows a substantial switch from oil 
to nuclear before the magic date of 2000A.D. One example of 
this is in the report of the Institute of Fuel's working 
party "Energy for'-t&e^Fiiture"' ^ 1) . OrQne. fofehthsi^oa^Qiii^jtfgns 
was that 

"Nuclear power generation must take over a large part 
of the increase in energy demand by 1980" 

In fact this conclusion is nicely ambiguous since it is not 

clear whether nuclear power generation will take over the 

supply of energy or itself contribofce..tec T -the increase in demand. 

A major aim of this paper is to show that it is extremely 
difficult to res61ve this ambiguity. Nuclear power programmes 
do consume enormous quantities of fuel, they do also produce 
large quantities of electricity. Until recently it was 
assumed that nuclear power stations represented a net source 
of energy. This assumption was, presumably, based on the 
fact that the stations could be operated at an economic profit. 



339 



3. 



However a nuclear power station is in economic competition 
with other electricity generating stations, and fossil 
fuel fired power stations are net consumers of energy. 
This means that if a nuclear system consumed 2kWh equivalent 
of oil, in mining, processing and transporting uranium, to 
produce lkWh of electricity it could still show a profit 
against a conventional station which consumes 3-4kWh 
of fuel per kWh of electricity.- In this case the nuclear 
system would be a net consumer of fuel and could be viewed 
as a better technology for producing electricity but not as 
a substitute for fossil fuel resources. In fact the 
situation is more complex than this. It is not possible to 
simply compare the energy inputs and outputs of nuclear 
power stations because the inputs and outputs occur at 
different times. Thus much of the following analysis is 
really a power analysis of nuclear programmes, not an energy 
analysis. This approach is described in detail in section 2. 
Section 3 documents the principle energy inputs and outputs of 
different types of nuclear reactors and describes the convention 
used in their evaluation. Section 4 then combines the method 
of power analysis with the energy values to explore the 
consequences of various policy decisions. 

2. Criterion for new energy sources 
2. 1 Energy profitability 

No one would suggest that building more fossil-fuelled electric 
power plants would alleviate a shortage of fossil fuels. It 
is understood that conventional power stations produce a 
secondary fuel by consuming primary fossil fuels. The 
distinction between primary and secondary fuels is based on 
the operation of the plant; in economic terms it is based on 
the current account. There is, however, a sense in which 
all fuel production is secondary production. This is based 
on the capital account of the industry involved. The 
investment of capital for exploration, drilling, machinery .and "so 
on, represents an investment of fuel which has to be made by 
allocating a proportion of present fuel supplies for the purpose. 
Fuels have to be consumed in order to make future supplies 
available. 



340 



Whilst the fraction of present fuel supplies allocated for 
future provision is small, say less than 5%, then the effects 
of this investment on the operation of an industrial economy 
are small. Clearly if ever the proportion invested rose to 
2 5 or 50% then there would be serious implications and energy 
policy decisions would have to take this into account. It 
is estimated that the present rate of fuel investment in the 
U.K is about 2.5% ( . (Note this is separate from the 27.5% 
of primary fuel supplies used in the operations of the fuel 
industries. The 27.5% is consumed in conversion and distribution 
processes and is on the current account of the industries 
involved.) However there are grounds for assuming that this 
proportion may rise to 10% or more in the forseeable future. 
This increase in energy investment and increased fuel 
consumption in operating the fuel industries means that in 
the future the gross energy requirements* (ger) , or energy 
cost, of fuels will change. This means that care is necessary 
in evaluating and interpreting the ger of fuels in the future. 
It also raises questions as to whether future fuel supplies 
should be discounted to present worth in a manner analogous 
to the discounting of future incomes in economics. 

The problem can be put into focus by considering a hypothetical 
new technology for producing fuel oil. Let us assume that at 
the moment the ger of fuel oil is 15000 kWht/ton. Let us 
further assume that the new technology requires an energy 
investment of 0.1 tons oil per ton produced as well as a fuel 
consumption of 0.1 tons oil per ton produced. Clearly the 
0.2 tons of oil needed to produce 1 ton of oil is part of the 
ger of the new fuel oil. But what ger should be attributed 
to the 0.2 tons of fuel oil? Should you use the present ger 
of fuel oil or should you set up simultaneous equations assuming 
that the process and investment oil consumed actually comes 
from the production plant itself? Throughout this report the 
convention used is that the investment energy should be counted 
with existing ger values whereas the process fuel for the new 

* This nomenclature conforms to that agreed at the international 
conference on energy analysis organised by IFIAS. Details of 
conventions available from M.Slesser; Dept. Chem. ,Eng. , 
Strathclyde University, Scotland. 



341 



technology should be deducted from the output of the plant 
when calculating the ger of the product. 

Dsing this convention it is clear that for an energy source 
to be energy profitable the net output over the lifetime 
of the plant must be greater than the gross energy input 
(i.e. the input energy taking account of the present ger of 
the fuels used) . 

A more stringent requirements would be that the total net 
energy output of the plant, discounted to present worth, 
should be greater than the gross energy input calculated 
at present worth. There are a number of reasons for 
questioning this approach. Firstly it is not obvious that 
fuel should be subject to a discount rate, there is no 
evidence to suggest that individuals or institutions place 
less value on 1 ton of oil next year than one ton of oil this 
year. It could be argued that technical progress will make 
more and more fuel sources available in the future and therefore 
any given stock of fuel resources is worth less in the future 
than today. Against this one can point to statements by oil 
sheiks to the effect that oil in the ground is the best type 
of investment they know. This reflects the fact that 
technical progress is a double-edged sword. New sources 
of fuel may become available but this may be offset by a 
greater demand for fuel or/and new uses for fuel feedstocks. 
For example chemical companies are coming to realise that 
petroleum may have a greater value as a chemical feedstock 
than as a process fuel. 

In principle it seems wrong to try to extend energy analysis 
into the realm of allocating resources over time since" eaergy 
is only one of many resources. It seems to us that economics 
is still the right tool to use for trying to allocate resources. 
What energy analysis can do is to tell you something about the 
physical consequences of different allocations thereby 
enhancing the accuracy of estimates of resource costs. 



342 



6. 



2. 2 Transient problems 

The distinction between energy investment and process 
energy is as important in energy analysis as is the 
distinction between capital costs and running costs in 
economics. The essential point concerning energy investments 
is that they must be made judiciously so that the proportion 
of total fuel supplies devoted to investments is kept small. 
This is especially true at a time when fuel supplies are 
tight. 

The problem can be clarified by shifting the emphasis from 
energy analysis to power analysis. This is illustrated in 
figure 1. Here the energy investment is represented by the 
area under the line, the energy output by the area above 
the line. According to the convention described in the 
previous section the input energy is calculated on the basis 
of the present ger of fuels and the output is a net output 
taking into account fuels used to drive the production 
process. To be energy profitable the area above the line 
must be greater than the area below the line. 

Figure 1 represents the power inputs and outputs for a «lngle 
unit in a new technology/ say the construction and operation 
of a nuclear power station. If a building programme involves 
constructing a number of reactors each year for some time 
then, as shown in figure 2, it may take a long time for the 
programme to show a "power profit". It will take even 
longer for the programme to show an energy profit. If each 
individual station shows an energy profit then the overall 
programme will also show a profit but not for some time. 
The time between initiation of the programme and on energy 
profit depends upon two parameters, namely , the ratio of input 
energy to ouput energy and the rate of building. Paradoxically 
in a time of fuel shortage it is important to build slowly. 
so as not to exacerbate the shortage! 



343 



«•) 




+ «- 



joMod ;ou 



■p 
u m 

0) p 
5 w 
O 

a p. 

e 

p s 

0) o 

c a 
c a) 

■H C 

o 

c 

O >w 
•H O 

p 
«J c 

-H O 
^ -H 
«J P 
> HJ 
M 
0) a) 

j: a 
h o 



344 



fa(fc) 




E -h 

O PU 

u-l II 
O 

treu 
c 

-H t^-P 
-P C -H 
-H -H >W 

3 go 
ui 3 u 



(0 >! 

M -" tr> 

d) u 
3 M 0) 

o <o c 
a a) d) 

>i 
c -p 

••HMO) 
Q) c 
c & 

O H • 
•h C O en 

(0 -H (0 

•H -p 0) Q) 

H (0 E >i 

(0 -P -H 

> w +j co 

vmoi 

M O E-" -H 



-5- « J9MOd J9U ► | 



345 



7. 



Most of this research report will be concerned with evaluating 
the magnitude and period of this transient in energy production 
associated with various nuclear power programmes. In section 3 
the energy inputs and outputs will be discussed and the time- 
scales e valuated in section 4. 



2 . 2 Long term problems 

There is a special problem associated with the energy analysis 
of nuclear power namely making allowance for fuel reprocessing 
and waste disposal. At the present time neither item represents 
a significant energy saving or energy consumption since both 
technologies are still being developed. Later in this research 
investigation we hope to be able to include reasonable estimates 
of the energy flows associated with these items. For the 
moment all we can do is point to their significance. 

The simplest way of illustrating the significance of these long 

term items is to redraw figure 1 on a much longer time-scale (figure 31 

The point is that nuclear wastes will have to be stored and 

perhaps maintained for thousands of years. The most important 

waste material is plutonium with a half-life of 24,000 years. 

Although plutonium is recovered from spent fuel rods 100% 

recovery is not economic (probably not energy profitable either) . 

Thus waste from spent fuel rods may have to be stored for 

240,000 years (10 half-lives). Now if a power station produces 

1000MW for 25 years but leaves waste materials which require 

a power input for maintenance of 100 kW for 250,000 years then 

the net energy output will be zero. A power input of lOOkW is 

equivalent to an annual consumption of 90 tons of steel or 

25 tons of rolled stainless steel. Hopefully designers of 

waste-handling will be able to keep the average power 

consumption below lOOkWas well as develop an infallible system 

capable of surviving for 250,000 years! 



346 



7W 



1 

V 



1% 

a e 



0) 4J 

£ (0 
•H « 
+> * 

M O 
S3 

si 



•H 0) 
CO 

I a 



^2S 



347 



3. Energy Analysis of Nuclear Reactors 

3. 1 Types of inputs 

At this stage in our research programme we think we have 
identified all the significant inputs to nuclear power stations 
but we only have approximate values for the ger of these inputs. 
Where possible we have indicated the accuracy of the values 
involved. Throughout the following sections all the data are 
in terms of inputs per 1000 MW installed capacity. In a later 
section we will show that the continuous average power output 
if a nominally 1000 MW station is much less than 1000MW. 

As pointed out earlier nuclear power stations are significantly 
more capital intensive than conventional stations. This is due 
to the greater complexity of a nuclear reactor as compared to 
a conventional oil or coal fired boiler and also to the need 
to assemble the entire core of the reactor before operations 
can be started. In order to make clear the differences between 
nuclear and conventional stations it is worth separating the 
capital costs into four types. 

In common with conventional power stations a nuclear station 
requires a significant investment in electrical machinery . 
This includes the generator set, power transformers, control 
and switch gear as well as distribution linka. Again common 
to both types of power station is the investment in buildings 
and services . This includes the site itself, office blocks, 
buildings to house equipment, cooling towers, service roads 
and the provision of an adequate water supply. 

The special requirements of nuclear power stations are then 
the initial core assembly and the nuclear reactor and steam 
system. The uranium fuel assembled in the initial core of a 
reactor represents about an eighth of the total fuel that the 
reactor will consume. For a coal fired station this would be 
equivalent to a situation where 14 mill-ion tons 'trt coal had to 
be mined and delivered to a 1000 MW coal station before it 
could be started up. As we shall show later the investment 



348 



of uranium in the intial core is a significant item in the 
energy analysis. The nuclear reactor itself, incorporating 
containment devices, safety and control systems as well as 
the steam circuits and heat exchanges also represents a 
significant input. 

Table 1 shows an approximate breakdown of the capital costs 
for nuclear and coal power stations. This data was provided 
by TNPG and is in good agreement with the estimates made 

previously on the basis of a rough nuclear cost breakdown 

(4) 

and a comparison of nuclear and coal stations. 

Table 1 . Estimated division of capital investment 
(£AW; 1973 prices) 





Coal 


Nuclear 


Electrical equipment 


50 


52 


Buildings and services 


25 


30 


Boiler 


17 




Nuclear steam system 




50 


Initial fuel 




14 


TOTALS 


92 


146 



From statistical surveys of the use of fuels in the U.K 
we know the average ratio of ger to £ value for all industries 
in 1968. These average values of energy requirement to financial 
value enable us to estimate the energy investment associated 
with the financial investments outlined above. However, such 
estimates are liable to serious error since the construction 
of a power station is a special undertaking and unlikely to be 
representative of an industrial average. This is especially 
true of nuclear reactors since these use special materials 
and special technologies. In pursuing this research programme 
our aim is to identify as many special inputs of materials 
and components as possible and to separately analyse these by 



349 



10. 
process analysis. After subtracting the financial values 
of these special inputs from the totals shown in Table 1 
we will be left with a residual which can legitimately be 
treated as an average industrial activity. 

To date we have performed a process analysis of the mining, 
production and enrichment of uranium fuel. In the near 
future we intend to perform process analyses on the following 
special inputs to nuclear reactors 

a) fuel cladding (zirconium) 

b) moderators and control rods (graphite etc.) 

c) heavy water 

d) pressure vessels 

3. 2 Energy values of inputs 

(i) Uranium fuel 

The uranium fuel input represents the largest single energy 
investment in the construction of a nuclear reactor. It is 
also the most problematic input from an energy analysis point 
of view since the ger of a fuel rod will depend upon the 
particular mine and the particular process technology involved 
as well as depending upon the degree of enrichment of the fuel. 
Figure 4a summarises the processes involved in providing an 

enriched fuel rod. The quantities of material shown on the 

(8) 

diagram are based on the average of a number of U.S mines 

using 0.3% ore and a stripping ratio of 20:1. Within the 
forseeable future much lower grade sources of uranium will 
have to be exploited if the proposed expansion in nuclear 
capacity takes place. For comparison purposes we have also 
analysed the processes proposed for mining uranium from 
Chattanooga shales with an average grade of C.007% U-0 g . As 
shown in figure 4b the quantities of material flowing changes 
dramatically with this change in ore grade. Tables 2 and 3 
summarise the energy analysis of uranium mining. 



350 



ii, 



s 




1 

•c 


► 


-9 




<-?• 




j: 













]? 




1 K 




5 3 




/ 


k 




i' 




5 


tf» 


o 


3 




J 




i 




* 




a 




i 


k 




j 




to 




r*- 


J? 




r> 




i 




j 


L 




jf 


1 




o 




c 


► 


* 





*i 



- I 



1 



■n 



J 



o i 



1 

2L 

3 

x 

X 






U 



351 



12 



■i 




1 




w 


j» 




& 




i 


k 




js; 












j. 




i? 




■3 •? 




5 M 




3 6 




i 


k 




I. 






*$ 


a? 




si 


3 




* 








i 




- 


k 




J! 


if 




h- 


SSi 




r» 


3 




'f 




2 




i 


i 




i< 






t* 


^^ 




i 




1? 






O 










S 





a 
I 



a? 



c c 

O 0) 

0) 3 

§ -2 






352 



13. 



There are a number of different processes used for milling 
processes and concentrating the uranium ore and of course 
the fuel consumed in the milling processes will depend upon 
the hardness of the particular ore-body. Thus in this 
process stage there is likely to be a wide variation in fuel 
consumption per ton for different mines. We have analysed a 
number of operations in detail and used the average fuel per 
ton processed. This average is 340 kWht/ton ore processed 
and comprises 110 kWht/ton direct fuel, 220 kWht/ton due to 
consumed chemicals, 6 kWht/ton due to machinery and a further 
4 kWht/ton due to the plant establishment. Combined with the 
average energy cost of mining 0.3% ore, shown as 336 kWht/ton 
ore in Table 2, this gives a total of 676 kWht/ton ore. This 
corresponds to 0.22 5 x 10 kWht per ton U,0 g . For comparison 
the values for the low grade uranium sources documented in 
Table 3 are 15.8 x 10 6 kWht/ton U-jOg and 12.5 x 10 6 kWht/ton 
U,0 R . Since 1 ton of U^O- contains 0.848 tons of uranium 
these results can be expressed in terms of tons and/ natural 
uranium. For the conventional mine the ger is 0.265 x 10 
kWht/ton U and for the Chattanooga shale mine 18.6 3x10 kWht/ton U , 

Table 4 documents the inputs required per ton of uranium for 
the conversion of tU0 fi to UF fi . Note that in this process 
analysis, and in all those following, the electrical input 
is kept separate and designated by (e) . All other inputs 
are converted to thermal equivalents and added together to 
give the (t) or kWht total. 

These processes, mining, milling, concentration and conversion 
represent the processes needed to make natural uranium available. 
The sum of these is referred to as the "natural uranium energy 
requirements" . 



353 



14. 



Table 2 



Energy analysis of Uranii 



'■ining 



Lucky 
Mc.mine 



kWht/ton ore 



Sateco 

mine 



Day- 
Loma 



Exploratory drilling 
Stripping overburden 
Ore mining 



10 

538 

33 



581 



5 

120 

55 

180 



10 
180 

55 



245 



Ore grade 

Stripping ratio 

Estimated reserves (10 tons ore) 



0. 31% 


0.21% 


0.3% 


35:1 


10:1 


18:1 


1.7 


0.7 


0.53 



Above data illustrates variation between different mines for 
energy inputs, ore grade and stripping ration-weighting the results 
in proportion to the estimated reserves give the following 
averages for five U.S mines. 



Energy input 
Strip ratio 
Ore grade 



336 kWht/ton ore 
23.7:1 
0.31% 



354 



Table 3. Mining and milling of low-grade uranium ores 



15 



(10 kWht per ton U 3 g ) 



Chattanooga shale 
(0.007%) 



Florida phosphates 
(0.013%) 



Mining 

Fuels 

Water supply 

Explosives 

Drill bits etc. 

Mine maintenance 

Establishment 

Mining total 



8.859 
0.370 
0.111 
0.089 
0.029 
0.043 

1.501 



0.430 
0.099 



0.008 



0.537 



Milling 




Fuels 


4.948 


Water supply 


5.425 


Replacement rods 


0.008 


Kerosine 


0.014 


Sodium carbonate 


0.020 


Other chemicals 


3.790 


Machinery (establishment) 


0.062 


Maintenance etc. 


0.010 


Milling total 


14.287 


Grand total 


15.788 



5.692 
0.327 

1.733 

4.118 
0.110 
0.040 

12.020 
12.557 



'3 W 8' 



355 



16. 



Table 4. Inputs for conversion U-0 g to UF g 



Inputs per ton uranium 10 kWh 

(e) (t) 

Water (24 x 10 4 gall @ 10 _2 kWht/g) - 0.002 

Natural gas (1.7 x 10 cu.ft. @ 
1035 Btu /cu.ft. 

Electricity 0.016 

Chemicals (HF @ 25 kWht/ton) 



0.052 



0.016(e) + 0.054(t) 



Table 5 . Natural uranium and SWU requirements per tonne of 

enriched fuel 

(based on 0.25% tails assay) 



Fuel enrichment Tonnes natural U Tonnes SWU 

per tonne fuel per tonne fuel 

1.6 2.94 1.25 

1.9 3.59 1.75 

2.1 4.02 2.05 

2.45 4.78 2.70 

2.6 5.11 3.0 
2.8 5.54 3.4 
3.0 5.98 3.8 
3.3 6.63 4.45 

3.7 7.50 5.30 
6.5 13.59 11.28 

10.0 21.20 19.08 



356 



17. 

The enrichment of uranium involves a partial separation 

2 38 2 35 
of the isotopes U and U . Natural uranium has about 

235 
0.71% U and enriched fuel rods have between 1.6% and 

235 
10% U depending upon the particular reactor design. 

The quantity of natural uranium required to produce one 

tonne of enriched fuel increases as the enrichment increases 



natural uranium (0.71%) gives x tonnes of enriched fuel (E%) 

and (1-x) tonnes of 'tails' or waste. Throughout this 

235 

report we have assumed a tails concentration of 0.25% U since 

(23) 
this is the value assumed by other authorities. Thus by 

conservation of U 

1 (te) x 0.71 (%) = x(te) x E(%) + (1-x) (te) xO. 25 (%) 

0.46 
Hence x = 1EHJ72 5) 



The tonnes of natural uranium required for one tonne of 
fuel enriched to E% is equal to /x. The values of /x 
are documented in Table 5 for various enrichments. Also 
shown are the "tonnes of separative work units" required 
per tonne of enriched fuel. The SWU's needed per tonne 
of fuel can also be deduced from a knowledge of the 
enrichment and tails concentrations although now 

the equations are more complex. There is a trade off 
between the SWU's required and the natural uranium required. 
The higher the tails concentration the smaller the SWU 
required but the more natural uranium is needed. Thus the 
numbers given in Table 5 are not absolute but reflect a 
particular compromise. (We have been told that current 
USAEC plants operate at 0.3% tails, but ref 24 states they 
operate at 0.2%) . 



The electrical input required per kg SWU have been given 
as 2.402 MWh/SWU (13) and 2.435 MWh/SWU (11) . We shall use 
a value of 2.42 + 0.02 mWh/SWU in this report. To this 
direct energy requirement should be added an appropriate 
proportion of the capital investment in the separation 
plant. £a more thorough analysis would incorporate the 



357 



18. 
proportion of the capital investment used by one nuclear 
power station in the energy invested in that station. For 
this analysis we will average the investment over all the 

fuel on an equal basis H The investment in an enrichment 

(12) 
plant is about £60,000 per SW tonne installed . Assuming 

that the plant lasts for 25 years this represents £2400 per SW tonne 

performed. Using the energy /value ratio appropriate to 'plant 

steelwork' this is equivalent to 96000 kWht/SW tonne. This is 

equivalent to 0.096x10 kWht/SW tonne. 

The final stage in fuel preparation is the conversion of the 
enriched (or natural) hexaflouride to uranium dioxide and 
assembling this into suitable fuel elements. Table 6 gives 
the inputs per tonne of uranium. 



Table 6 Inputs for fuel conversion and fabrication 

Per ton of uranium 10 kWh 

(e) (t) 

Water (1.48 x 10 5 gall @ 10~ 2 0.001 

Natural gas (1.02 x 10 5 cu.ft. 

@ 1035 Btu/cu.ft. 0.031 

Electricity 0.048 



Total 0.048(e) 0.032(t) 



Note that the procedure adopted gives a larger energy ratio 
since (E -4e)/(E. +Ae) is smaller than E /E . . 



358 



19. 
These results can be brought together to give three components 
for the energy required to produce a tonne of reactor fuel. 
These components are the energy required for natural uranium, 
for enrichment and for fuel fabrication. These are set out 
in Table 7 and include data based on present ores (0.3%) and 
future ores (0.007%). In order to calculate the energy invested 
in the core of a nuclear reactor the electrical and thermal 
components of energy required must be combined on the basis of 
the present conversion efficiency for electricity. These are 
the values given in brackets at 25% conversion. 

In order to calculate the energy required for different reactors 
we need three pieces of information. For a given reactor we 
require its net electrical output, its initial fuel inventory 
and the fuel enrichment. Herein lies a problem. Before revising 
this report we had too little data to make a sensible 
evaluation, now we have too much data. It is too much because 
no two authorities can agree upon the values to be given to 

these parameters. We have ended up using the data given in 

(14) 
the Directory of Power Reactors when it is consistent with 

that given in the Nuclear Power Index. These data sources 

give details of actual reactors and whilst these seem to be 

reliable the reactors may not be characteristic. By and large 

we have tried to choose reactors which are close to the 

(23) 
characteristics given by OECD or by private communication 

from UKAEA and TNPG. However this has not always proved 

possible. 

For Magnox we have used ' Oldbury A* which is within 10% of 
OECD and UKAEA data. For SGHWR we have used the design 
published by Moore et al which is consistent with other 
data. For the PWR we have used four reactors since there 
is an enormous variation in the parameters from one station 
to the other. All the PWR's used are large ( > 500 MW) 
stations of modern design. Data listed under (i) refers 
to Haddam Neck (ii) refers to Maine Yankee (iii) to Jos M.Farley 
and (iv) to Shearon Harris. We have not been able to find 



359 



20. 

any actual PWR station with the 0ECD v ' or UKAEA characteristics, 
For the AGR we have used Hunterston B and for CANDU Pickering. 
The data on the HTR is from TNPG. 



Where an initial core has a distribution of fuel enrichment 
we have evaluated the weighted average enrichment and used 
this to calculate the energy requirements. 



(ii) Heavy water 

The SGHWR and CANDU reactor systems make use of heavy water 
(D 2 0) as a moderator. This represents another significant 
energy investment. In financial terms the heavy water 
inventory of 1000 MW station costs about £9m. It is known 
that direct fuel costs contribute slightly over 55% of the 
total cost of heavy water (16) . Since the kWht/£ of fuel 
is now about 1000 then the kWht/£ value of heavy water should 
be about 550 kWht/£ . On this basis the heavy water represents 
an energy investment of 4950 x 10 kWht. This value is only 
ai approximate estimate and has a probable error of + 20%, 
furthermore it is based on the operating costs of the electrolysis 
process. 

The Heavy Water Division of Atomic Energy Canada have indicated 
that their modern plants (based on H.S process) have much 
lower fuel costs. They estimate that the energy inputs per 

tonne of D-0 are 0.65 x 10 6 kWh (e) plus 6 x 10 6 kWht (t) of 

(22) 
steam . Since this is an inventory item the electricity 

is charged at present generating costs giving an energy 

requirement of 8.6 x 10 kWht/tonne DO. The CANDU reactor 

uses heavy water both as a moderator (0. 3te/MW (e) ) and as 

(27) 
a coolant (0. 4te/MW(e) ) with a total inventory of 0.7te/MW(e) K ' . 

The SGHWR reactor uses heavy water as a moderator and has 

an inventory of about 0.25 te/MW(e) ^ ' . 



360 



TABLE 7 Energy costs of uranium operations 
(10 6 KWh) 



20(a) 



Mining and milling 

Conversion 

Total per tonne U 

(^ 25% conversion) 
(for refuelling) 

Enrichment 



0.3% ore 


0.007% ore 


0.265(t) 
0.016 + 0.054-(t) 


18.63(t) 
0.016(e) + 0.054(t) 


0.016 + 0.319(t) 


0.016 + 18.684(t) 


(0.383(t)) 
(0.335) 


(18.75(t)) 
(18.70) 



per SWU 



2.42(e) + 0.096(t) 



(<S 25%) 

(for refuelling) 



(9-78) 
(2.52) 



J''uel rod fabrication 
(per te fuel) 



0.048(e) + 0.032(t) 



(«3 25%) 

( for refuelling) 



(0.224) 
(0.080) 



361 



■ 

>H 

EH 


ft 

O 
O 

o 


C\J 

d; 

CO 


en 

CT> 
CM 

m 


T 
CD 

CN 


ft 

o 

r<A 
H 


•-1 




CAJ 
CAJ 


KA 
UA 


oo 

CN 
00 


NA 


o> 


en 
oo 


CN 

O 
VD 


CN 


W 


UA 
H 


UA 
UA 


^ 


<N 
vO 
CN 


O 


• 


tf\ 


CAJ 


IN 


<JN 


UA 


r— 1 


EH 


o 






K> 


:* 


CAJ 


UA 




a 




















o 




















•H 




















CO 


CO 


\0 


ON 


CO 


CAJ 


C^ 


1 


I— 1 


UA 


Sfc 


Oj 


KA 


CAJ 


CAJ 


CAJ 


1-1 


d" 




H h h 




















CD X> CD 




















co « 




















(=4 Ph CD 




















f-l 

0) 


o 


1 


W5 


Pi 

VO 

VD 


U0 


CN 
US 
CN 
CN 

r-l 


00 


UA 
CN 


CAJ 


in 


a 

CD 


O 

• Ph 


oj 
co 


O 

CN 


ON 


CO 


CN 


^ 


m 




O O 


i— ( 












•-{ 






p 

-p 
CO 






















ro a> 


KA 


CN 


c 
m 
m 


MA 


o 
in 

CN 


cn 


CN 
UA 


O 
IN 


oC 

i-H 
i-H 


s 


O O 




















3 ft 

•H Ph 




o 

CN 


m 
m 


UA 


s 


ON 


£ 




3 

o 


Ph CD 

a c: 


1 


ft 


3 


CN 

OJ 


UA 


' 


UA 

CAJ 


0) 0) 




















a><~» 




















S oco 




CO 
CN 
CI 


oo 

LTI 


CAJ 


U3 


00 


CN 




U0 


02 




i— 1 


UA 


CS 


OJ 




UA 


Ph CD 




K\ 


d- 


CAJ 


UA 




OJ 


<1> CD -P 




















-P S^w 




















-o 




























































atur 
rani 
equi 
te) 


CN 




CN 
00 


hO 
ITS 


m 




C\J 
K\ 
OA 


C\J 
CO 


00 

o 
ro 


S3 3 Sh^ 




















X! 










UA 




UA 






o 




,— | 


hO 


vD 


hO 


CN 


,^- 




UA 


•H^. 


1 


• 


• 




• 




• 


1 


• 


u 




CAJ 


K\ 


OJ 


r<A 


OJ 


OJ 




uD 


a 




















(D 




















>a 




















P 




















O 




















-P 


r<A 


8 


ft 


d- 


CN 


CN 


UA 


OJ 


CN 


a 


CN 


O 


QN 


CO 


OA 


CO 


• 


r-H CD^-N 


OA 
















OJ 


<D > CD 


















OJ 


3 a -p 




















pc^-Hv^ 


























/ ^ N 


^^ 


^ 


^ 














■H 


•H 


•H 


> 














^ 


•H 


C2 


v— / 










X! 




















O 














p 






S 














Q 






o 




« 








« 


i 


On 




a 


o 


3 








O 


Eh 






CO 


Pm 








<a! 


o 


W 



20(b) 



362 



21. 

(iii) Other inputs 

The remaining inputs are those documented in financial terms 
in Table 1 under the headings 'electrical equipment' , 
buildings and services' and 'nuclear steam system'. Since 
no detailed data is available the best that can be done is 
to use the industrial average kWht/E value deduced from the 
1968 Census analysis and corrected for changes in the £ value 
(i.e. allowing for inflation). This gives values of 33 kWht/E 
for electrical machinery; 30kWht/£ for nuclear steam system 
and 2 8kWht/£ for building and services. Using these multipliers 
and the approximate cost breakdowns given by the CEGB for 
different reactor types the approximate energy inputs can 
be evaluated. This is summarised in Table 9. For reactors 
using enriched fuel the initial fuel is a major item in the 
total energy invested and since the uranium ger is known to + 
10% the overall totals are probably accurate to + 15%. For the 
Magnox reactor a large proportion of the total energy requirement 
has been deduced from financial data and here the error is 
probably + 20%. Thus the final totals do indicate the order 
of magnitude of the energy investment but are not accurate 
enough to permit detailed comparisons of different types of 
reactor. 



363 





ft 


C\J 

o 


•>* 


o 
o 


vO 
00 


O 


A 
tN 


CA 
CA 


ES 

vO 


o 


rH 

o 


8 


IN 

A 


o 

CM 


0> 

in 


R 


o 
CO 


^ 


CO 

cs 


o 
3- 


vo 


3 


o 


CM 


fM 


CM 


1— 1 


R 


r-i 


OJ 




ft 


rA 


r-i 




00 
VO 


rA 
rA 


00 


r-i 
LN 


r-i 
00 


LA 
OJ 


d 


EH 


• 


00 


CN 


O 


ro 


00 


C^ 


IS 


CN 


00 


g 


O 


IN 


o 

i-H 


O 


CN 


CO 


vD 


o 


O 

r-i 


0> 





ft 


C\J 
vD 


cr> 


*tf 


ft 




0> 

rH 


rA 


PA 

A 


• 


O 


o 
o 


$ 


CM 


00 


ft 


VO 


it 

r-\ 


VO 
OJ 


lA 


00 


O 


o 




i— i 


CM 


rH 


H 




OJ 




00 


3 

-P 






















^. 


, 


c* 


CM 


cs 


OJ 


LA 


LTN 


r-i 


r* 


•H 


S> 


0^ 


00 




O- 


Ml 


rH 


LA 




<N 


d 


i 


a 


^ 


O 


00 


r>- 


0-N 


A 


r-i 


VD 


H 


o 




CO 


vo 


PA 


v 


OJ 


A 




CM 


ni -p 






ft 












8 




a) ctf 






rH 












s 




W^ 






OJ 
















a 


vo 

r-i 


IN 
vo 


ft 


ft 


ft 


o 

A 


CA 

00 


CN 
VD 


s 


CO 






















CO 


-p 




















CO 
!3 


S 


o 


o 


O 


O 


O 


o 


O 


O 


O 




M 
vD 


CO 


rH 
St 


rH 


s 


ft 

rH 


r-i 


IN 
vO 
OJ 


H 
OJ 


8 

rH 




O 












































S 


rA 
CN 


rH 

rA 


o 

IA 


ft 


R 


ft 


ft 


rA 


ft 


n 






















d 










































•H 


4-5 




















T3 


| 


8 


00 


o 


Q 


Q 


o 


o 


CO 


Q 


•H O 

2 -P 

PQ 


vO 


VD 

CO 


& 


a 


a 


£ 


£ 


vD 

00 


1 




O 












































a 


C\J 


C\J 


OJ 


OJ 


OJ 


OJ 


OJ 


OJ 


OJ 




<H 


A 


a 


ITS 


LTN 


LT\ 


LT\ 


a 


A 


A 


o 










































•H 


p 




















-p 


1 


VO 


vO 


VO 


vO 


vO 


VO 


vO 


vO 


vO 


O 


^ 




rH 


r-{ 


rH 


rH 


rH 


r-i 


rH 


rH 


(D 




O- 


IN 


CN 


CS 


CN 


CN 


IN 


CN 


(N 


rH 


03 




rH 


rH 


rH 


rH 


rH 


r-i 


rH 


rH 


w 


O 




























^ 


<r> 


^ 


^ 
















•H 


■ H 


■H 


> 


















•H 


5 


•H 












XI 






















o 














S3 








a 














P 








1 


o 


S 








cj 


9 


| 






CO 


& 








<JJ 


o 


a 



21(a) 



364 



22, 



3. 3 The net outputs 

To calculate the net energy output of a power station it is 
necessary to take into account 

a) station lifetime 

b) station load-factor 

c) distribution losses 

d) use of electricity by CEGB 

e) energy to replenish fuels 

Of these factors the most important are the first two, station 
life-time and load-factor. These are also the most problematic. 
So far no commercial nuclear power plant has had time to operate 
for its design life and possibly because this is a new 
technology, only a few stations have achieved anything close 
to their design load-factor. Even when these teething problems 
are over there are still a number of uncertainties which make 
it difficult to arrive at firm data. The CEGB operates all 
its power stations on a 'merit order' so that at any time 
the demand for electricity is being met by the most economical 
stations. At the moment the best nuclear stations are close 
to the top of the merit order and consequently have a high 
load-factor. However as technology develops and new stations are 
built then these stations will be demoted in the merit order 

and will have a lower average load-factor simply because 

(4) 
newer stations can be run cheaper. Thus it has been estimated 

that over a design life of 25 years a nuclear station with an 

initial load factor of 70% (when at the top of the merit 

order) will have an average load factor of 62%. When compared 

to actual operating experience of nuclear reactors it is found 

that only a few achieve their 70% load factor early in their 

life . Upto 31st December 1969 only 10 commerical reactors 

had achieved load-factors of 70% or more. After allowing for 

a derating Magnox reactors achieved an average load factor 

of 65% 1972 - 73* 18 ^. The data for large 600 MWe PWR reactors 



*The USAEC uses a life average load factor of 57% for its 
projections 



365 



22(a) 





B 






c* 


U3 


VO 


O 


ON 


r» 


H 




o 


in 


e* 


^r 


CS 


in 


8 


CO 


m 


CN 




o 


o 


«H 


rH 


IA 


CO 


& 


CN 


O 






in 


in 


** 


00 


CN 


OO 


. i-t 


CO 


3 


d 


m 


CN 




































o 


ft 


r- 


ia 


H 


| 


vl/T 


kO 


0> 


r» 


m 






r^ 


c^ 


K> 


00 




CN 


CN 


r* 




d 




st 


C\J 


CM 


00 


00 




m 


a 




















o 




















■H 




















P 




















CO 




















3 b 


in 


C\J 


C\J 


oo 


CO 


OJ 


OO 


m 


UA 


h h 


















O 


,13 CD 




















cO d 




















p 




















d 




















CD 




















, 


VD 


s 


CJN 




CO 

CN 


OO 


, 


tn 

CN 


d-H U 




H 


tA 


r-\ 


cm 


H 


00 




ft 


a u cu 




















lid 




















H 5 
CO CO 






















T 


CA 


<?> 


TT 


Lf\ 


& 




ON 


d 


1 


vo 


5* 


(N 


a\ 


t 


1 


00 


dpi) 

§5 -p 












O 






<-{ 












CN 








<j w^ 




















13 


ft 


o 


»x> 


r- 


ITS 


CD 


OO 


rH 


CN 


V£> 


h 


O 


0> 


in 


i CO 


CN 


1-t 


<D 


in 


cn 


CD 


o 


^" 


ro 


r» 


res 


vO 


oo 


CO 


CN 


vo 


1 


d 


, oo 


CN 


m 


OvJ 


CM 


OO 


.i-H 


CM 






















, 


fc 


CN 


CN 


0^ 


rA 


r> 


xD 


H 


00 


C^ 


-P 


• 


^D 


*r 


tf) 


^ 


T 


ICS 


LT\ 


OO 


st 


CO 


o 




















& 






















-h a a 
co 3 <u 




















,t> 


vo 


in 


CN 




00 


ITS 


CN 


CM 


-P CO P, P 


J*> 


CN 


4 


OJ 


O 


o 


LA 


<D 


■t 


"H 


i-H i 


CM 




"S» 


H 


r-i 




■** 






o 




i-H 








V 


CO P, CD 






CN 














S l=> ft 




















^ 


Pq 




















CD 


1-5 


r- 


cj- 


(N 


OvJ 


<H 


St 


R 


CN 


CO 


P 




oo 




C\l 


OO 


<N 




V0 




~^-s 


» 


i-i 










CN 






VO 


H 


vO 




















CD 

d 










































<H 


Ph 

3 






g 


8 


rA 




9 

10 






cO 




i—{ 


>x» 


IA • 


ir\ • 


O • 


CO • 


CO 


o° 


d 
d 


§ 


o 


o 


rA 


KNOJ 


.rA 
rA@ 


C0^ 


4" 00 


o 

H 


| 


o 




uS 


4-® 


® 


O0@) 


® 




^d, 


-3 


H 






~*-^ 


^ 


v^ 


^ 


V' 












r ^ > 


•-N 


C-N 


^ 














■H 


•H 


•H 


> 
















•H 














M 






s_> 


■H 


v_/ 










o 














E> 






& 


5 












Q 






3 


o 


n 










9 


m 

EH 






CO 


* 








<3j 


O 


a 



366 



23. 

in the U.S. A shows a variation of load factor between 23.3% 
and 93% with the average below 60%. Thus although the 
initial design load factor is 70% it seems that the time 
average 62% is more realistic even in the short term (and 
it is the short term load factor which is significant for 
the energy analysis of the transient) . 

Throughout the discussions of nuclear reactors it is 
assumed by engineers, operators and accountants that the 
plant life is 25 years. No reasons are given for this and 
there is no indication of whether this is an economic life- 
time or a technology limited life-time. For the moment we 
will concur with the prevelent assumptions and also use a 
lifetime of 25 years. 

In the past the distribution losses and use of electricity 
by the CEGB has accounted for 15% of all electricity generated 
Of this 7.5% is classed as distribution losses and 7.5% as 
"use by CEGB and area boards in offices, showrooms, workshops 
etc.". Now the data for nuclear power plants refers to costs 
per " kW sent out " and thus have already subtracted power 
station uses of electricity. The use of electricity in CEGB 
offices and Area Board showrooms etc. has not been taken into 
account. We shall assume that the power station use represents 
3.75% of generated electricity and the use by CEGB etc. the 
other 3.75%. Thus for nuclear stations the total loss incurred 
from power station output to final consumer is 11.25%. 

The final item to be considered is the refuelling energy use. 
To calculate this we need to know 

a) the annual refuelling in tonnes uranium 

b) the source of uranium and type (e.g. enriched). 

This data can then be combined with that shown in Table 6 to 
calculate the energy used in this activity. Note that for 
this purpose it is wrong to convert the electrical input in 
the ger of uranium to a coal equivalent. We want to calculate 
the net output of an electricity station so we simply subtract 
the electricity used for producing uranium from the station 
output and do not introduce any thermal efficiency factor. 



(20) 



367 



24. 

The final value of the energy ratio for a particular reactor 
is very sensitive to the refuelling tonnage. This is because 
this item subtracts from the power output and over 25 years 
a small change in power output gives a large change in energy 
output. The refuelling data appears to be given on the basis 
of 100% load factors. We have therefore corrected these 
refuelling tonnages to our 62% load factors. In reactors with 
distributed cores we have assumed that the refuelling is at 
the highest grade of enrichment (since partially spend fuel 
rods are moved out of the high enriched zone to lower enriched 
zones) . We cannot make sense of the AGR refuelling data which 
is quoted at 100% enrichment. This degree of enrichment would 
require an incredible amount of SWU. In the absence of other 
information we have taken the highest enrichment of the initial 
core. 

Table 10 gives the annual energy consumption for refuelling for 
the same stations as documented in Tables 7-9. Table 11 calculates 
the real energy output, over a 25 year period, for each type 
of station. It is important to note that the lions share in 
the reduction from a nominal 1000MW(e) output is due to the 
62% load factor. This represents electricity not generated and 
is not a loss factor. The heavy water reactors have their 

net output adjusted to make allowance for annual loss of heavy 

f 22) 

water. For CANDU this is 0.7%/annum ' corresponding to a 

power loss of 4.80 MW. For the SGHWR the loss is about 0.5%/ 

(2 8) 
annum corresponding to a power loss of 3.43 MW» 



Note that this procedure does not include the fuel used in 
generating electricity used by the station internally. This 
is about 5% of gross output, so our refuelling tonnages are 
too low by 5%. 



368 



25. 
3.4 Energy ratios for different reacto r s 

The energy ratios for different reactors can be deduced by 
dividing the energy outputs shown in Table 11, by the energy 
investments, shown in Table 9. The results are given in 
Table 12. However some care is needed in interpreting the 
crude ratios. 

The errors shown in Table 12 are those calculated on the basis 
of our estimates of the accuracy of our energy evaluations. 
The indicated errors thus assume that the reactor parameters 
used are absolutely correct. In view of the contradictory 
information that we have received since publication of our 
first report this is an unreasonable assumption. However 
we have absolutely no way of assessing the validity of 
reactor parameters. It is an indication of the sensitivity 
of our analysis to these parameters that PWR (i) is the worst 
reactor on 0.3% ore whereas PWR (iv) is the best reactor on 
0.3% ore. 

To some degree the uncertainty in reactor parameters must 
be due to differences in assumptions used when quoting or 
calculating the parameters. We feel that there is an obvious 
need to obtain real operating data from a wide range of 
stations and use this actual data to evaluate energy ratios. 
This would also more accurately reflect the differences in 
actual performance of the reactors since (we have been told) 
only a few reactors live up to their design parameters. 



369 



25(a) 



; : 





d 






















•H • 


















o 




03 


in 




o 


on 


in 


in 


o 


O 


CO 




& £ 


o 


r» 


co 


rH 


o 


o 


.oo 


CO 


ON 




<n 


r- 


r- 






in 


CN 


ON 


** 




rl 


(N 


VO 


l£> 


V£> 


ON 


in 


CO 




•^ 




lA 
1 ^ 


co 


m 




m 




l£> 


** 






QJ 


-P 




















u 





in 


















o 


p 


CN 


CN 


VD 


vO 


CN 


LA 
OJ 


vo 


5 


o 


R 


g s 


o 
in 


on 
m 




LA 

OJ 




CS"N 
ON 


ON 


in 
o 


o 






CN 






CN 






CN 


o 














OJ 










Cfl 




















o 

























bO 






















a 

■H 


O 
o 


oo 

00 


r~ 


£ 


CM 


r-{ 
LA 


co 
m 


1 


ON 




« hS 


*r 


CN 




(\J 


CO 


OJ 


CO 


r-i 


CO 




3 <D 






















«H ? 






















<D O 






















«CL, 






















d 




LA 


<^ 


lA 


r- 

00 


in 
o 


Q 
ro 


o 


o 




w 


oo 


0J 


OJ 


2 


00 

co 


in 


in 

ro 


© 


00 




& £ 


in 


tf 


0"N 


;$■ 








CO 


i-H 




00 


H 


o 


rH 








rH 






ri 




rH 


rH 










rH 






LA 


^ 




















1 ^ 






















P 




















<a 


d 




















u 


a< 














LA 






o 


-p 


*r 


rA 


R 


3" 

OJ 


oo 


LA 
LA 


CO* 


00 
rH 




Fr\ 


rH 


m 


OJ 
lA 




OJ 
lA 


in 


OJ 
LA 


lA 


L^ 


o 
in 


o 
























faD 






















d 






















•rl 


00 


3" 


oj 


CO 


m 


L>- 


CO 


OJ 


VD 




« hi 


CO 


rA 


o> 


CN 


CN 


=t 


rl 


rA 


A 




d 




oj 


<t- 


OJ 




OJ 


rA 








<H > 






















o 






















« Pm 




















d 
> o 
























LA 




lA 


LA 


lA 


LA 


LA 


LA 


LA 


■rl d 




Oj 


00 


C\J 


OJ 


OJ 


OJ 


OJ 


L^ 


OJ 


p a p 


g 


















O-H d 
<D -P d, 


o 

LA 




R 


o 

LA 


£ 


R 


LA 




£ 


=h d p 




Lf\ 


LT\ 


A 


LA 


A 


LA 




LA 


tHO? 






















w oo 






























^ 


^ 


^ 


^~N 
















•rl 


■rl 


•rl 


> 




















•rl 














X 






















o 


P-< 






v_> 






!=> 








s 














« 








5 

i 


o 


S 








a 


^ 


EH 






CO 


Pm 








< 


o 


W 



370 



26. 



Table 12 



Energy ratios for nuclear reactors 



Magnox 
SGHWR 
PWR (i) 

(ii) 
(iii) 

(iv) 

AGR 

CANDU 
HTR 



Uranium from 


0.3% ore 


0.007% ore 


15.1 + 3 


1.28 + 0.2 


11.2 + 2 


2.56 + °- 4 


10.9 + 2 


0.65 '+ 0.1 


15.6 + 3 


3.28+ 0.6 



12.9 + 2 
16.5 + 3 

10.5 + 2 
11.1 + 2 
15.8 + 3 



2.35+ 0.4 
4.23+ 0.8 

1.56+ 0.3 
6.24+ 1.2 
3.56 +0.6 



371 



27. 
As mentioned earlier the most sensitive parameter in our 
analysis is the annual refuelling tonnage. This is also the 
parameter which is likely to depart most from design data 
since the largest uncertainties seem to be in estimating fuel 
burn-up rates and overall thermal efficiencies. To indicate 
the sensitivity to this parameter we have also evaluated the 
energy ratio for PWR (iv) assuming that it uses 10% more fuel 
per annum thatn the quoted value of 28.0 te. For 0.3% ore 
the new energy ratio is 16.4 which is only 0.6% different from the 
value in Table 12. However for the 0.00 7% ore the new energy 
ratio is 3.54 which is 16% less than the value in Table 12. The 
sensitivity can be identified on Table 11 where the refuelling 
power for the 0.00 7% ore becomes comparable to the equivalent 
continuous output. The difference between these two comparable 
numbers amplifies an error in either. 

There are a number of obvious conclusions which come out of 

this analysis of different reactors. At the very low ore grades 

the reactors which show up best are those which use the 

smallest quantities of natural uranium, this puts CANDU at 

a clear advantage. The large heavy water inventory of CANDU 

puts it at a relative disadvantage for the low grade ore. 

If the natural uranium requirement were to become important, 

either because of resource shortages or because of steeply 

rising fuel costs, then the enrichment process could be 

changed to give a tail assay of 0.2%. This would decrease 

the natural uranium required for the enriched reactors, but 

increase the SWU requirements. These effects tend to cancel 

in energy terms but the overall saving in natural uranium 

could delay the time when very low grade ores have to be used. 

For comparison Table 13 shows the overall ratios for three 

cases. The first case corresponds to 0.2 5% tail assay throughout 

and repeats the data on Table 12. The second case assumes 

the reactor is built using a 0.25% tail assay but is refuelled 

using 0.2% tail. The third case assumes that 0.2% tail assay 

applies throughout, i.e. to the initial core and refuelling. 

Although all the reactors show an improvement the only significant 

change is for the HTR. (All the other changes are within 

the error estimate for the 0.25% tail case). The HTR improves 



372 



28. 



most because it uses highly enriched fuel and changing 
the tails assay has a significant effect on the natural 
uranium requirement. 

To significantly improve the energy ratios of the other 
reactors using enriched fuels it is necessary to develop 
energy cheaper mining methods. Energy cheaper methods of 
enrichment (for example the centrifuge system) would not 
substantially alter the energy ratio for low grade ores 
since 90% of the refuelling energy and more than 70% 
of the initial investment is expended in mining operations. 



Table 13 Energy ratios for low grade ore 
(all uranium from 0.007%ore) 



Magnox 



PWR 



AGR 



CANDU 



HTR 





0.25% 
tails 


0.2% 
tails 
for refuelling 


0.2% 
tails 

initial and 
refuelling 




1.28 


+ 


0.2 








2.55 


+ 


0.4 


2.74 


2.81 


(i) 


0.65 


+ 


0.1 


0.90 


0.96 


(ii) 


3.28 


+ 


0.6 


3.51 


3.58 


(iii) 


2,35 


+ 


0.4 


2.74 


2.88 


(iv) 


4.23 


+ 


0.8 


4.46 


4.55 




1.56 


+ 


0.3 


1.73 


1.78 




6.24 


+ 


1.2 








3.56 


+ 


0.6 


5.45 


5.64 



373 



29 



4. Power Analysis of Building Programmes 

4 . 1 Principles of analysis 

Before embarking on the detailed evaluation of the energy inputs 
and outputs of particular programmes it is worth elaborating 
on the description of the transient analysis in section 2.2 so 
as to clarify the principles involved. Figure 1 illustrates 
the basic power graph for a single nuclear reactor. Throughout 
this evaluation the base case will assume that the construction 
time T c , is 5 years, the delay before power output (T ) 1 year 
and the station lifetime, T , 25 years. Later the consequences 
of varying these time parameters will be discussed. For 
simplicity it is assumed that the total energy input can be 
represented by a constant power input over the construction 
period. This means that we are assuming that in this five year 
period all the uranium, steel, concrete, copper etc. incorporated 
in the reactor is produced from raw materials. (Later we 
plan to try to construct a more accurate power graph and to 
investigate the effects of different profiles on the conclusions.) 

In the previous section it was shown that the ratio of E ^./E. 
ranged from 16:1 to 0.6:1. To illustrate our procedures we 
will use a ratio of 10:1. 
since 

E in = P : 

E = P 
out ( 



E 
o 

E i 



Now let us use this ratio to construct the graph shown in figure 
2 which is based on starting one reactor per year. For the 
first five years of this programme the power input steadily 
increases to 5P . . In the sixth year work on the first reactor 
stops 



374 



30. 



on the second reactor stops and work on the seventh reactor 
starts, so the input is still 5P.. However in this year the 
first reactor comes on line so the net input is (5P.-P ). 
In the eight year the net input is 5P.-2P , in the ninth year 

5P.-3P and so on. Given that P =2P . the point of zero 

1 o o 1 c 

power input is when the net input is (5P.-2.5P ), i.e. eight 
and a half years after starting the programme. The total area 
under the graph is approximately equal to 19P. (by calculating 
area of triangle) . To retrieve this energy input the area of 
the triangle above the line must equal 19P. . Since the slope 
of the power line is such that its height is 2P.xT where T 
is the time we have 

h (2P i .T) x T = 19 P i 

so that 

T 2i 4.5 years 

Hence for this example the total time for net energy profit is 
13 years and the time for zero power input 8.5 years.* 

This simple calculation illustrates the significance of the time 
parameters, energy ratio and building rate on the times for 
energy profitability. The detailed calculations based on the 
data in Tables 8 and 9 were performed -using- a sipple computer 
program. The building program is specified by a function N(T) 
which gives the number of reactors started in each year (T) . 
Then the number under construction 1 is given by n (T) where 

n c (T) " 21 XT) 
T-T c 

and the number of reactors on line at time T is given by n f (T) 

V 

XT) 



where T- (T +T ) 



2 



The net power output of the programme is then given by P(T) where 
P(T) = P n.(T) - P.n (T) 



*Note that in the NEL paper, ref. 21, this simple calculation was 
mishandled and an incorrect time gf 19.5 years for energy profit 
deduced. 



375 



31. 



(**-<g|>H 



The total energy input is given by E (T) where 
T 
E(T) = £ p(T) 

Thus the outputs of the programme are the time series P(T) 
and E (T) for inputs of P . , P , T , T and the function N(T). 

When the building schedule is a simple analytic function 
the above equations can be solved by replacing the summations 
by integrals. The- the number of reactors under construction, 
h (T) , is given by 

T 

(t) dt (for T<5] 



dt (for T>5) 



n c (T) - f n. 
n c (T) = \ n i (t) 

c 

similarly n f (T) - \ n ± (t) dt (for T>6) 

The expression for P (T) remains the same but the evaluation 

of E (T) requires integration taking account of the discontinuities 

in P (T) . Thus 

f 6 C C 

E(T) = P q \ n f (t) dt - P i \ n ± {t) dt-P i \ n^(t)dt 



376 



32. 



4. 2 Results for various building schedules 

In view of the uncertainties in the results derived in section 
3 it would be wrong to go into details of comparing different 
reactors constructed on different schedules. Instead we will 
point to a number of important general conclusions which can 
later be combined with accurate data so as to evaluate various 
nuclear policies. Most of the calculations that follow have been 
done on a discrete basis, i.e. starting an integral number of 
reactors each year. This gives stepwise variation in the power 
and energy variations. Where this does not confuse the diagram 
we have shown the discrete steps. However in a number of 
diagrams we have had to use smooth curves, which to a degree, 
are misrepresentative. 

The first investigation was to examine the relationship between 
the time for power and energy balance and the energy ratio. 
This is shown in figure 6 for a building schedule of one 
reactor per year. Here the output power of each curve is the 
same and the input power varies with the energy ratio. Thus 
after year 6 all the lines have the same positive slope, but 
they all have different negative slopes in years 1-5. The 
curve labelled with an energy ratio of 10 corresponds to that 
shown in figure 2 and discussed in section 4.1 The corresponding 
energy graphs are shown in figure 6. Here the variation is 
shown stepwise to emphasise the discrete variations involved. 
The variation in time for energy profit and the energy ratio (E_) 
is shown in figure 7. This shows that even with E in the range 
20-25 the time between initiation and profitability is close to 
10 years. When E falls below 10 then there is a steep increase 
in the time until when the ratio is 2.5 the time exceeds the 
nominal life time of the first reactor. With this type of 
building schedule it is obviously desireable to have E R greater 
than 10. 



377 



33. 




a (J 
Mi C 

o o o 

>w J=. 

M T3 
CO — - 0> 

x: to 
Cum to 
m o xi 

u -H 

(0 0)03 

M M c o 

(1) -H >, 

O fcn ^ 

a u x o> 

0) U Cu 
T3 G <C 

Q) Q) 0) C 
O 

-P <4-t -H 

C O -P 
0) fd 

0> C W 
(4-1 0) 

<U <D 

•H .p C 

•o to o 



+ ♦— JdMOd jau » | 



U O - 76 - 25 



378 



34 



net energy 




Figure 6. The energy graphs corresponding to figure 5. 



379 



30- 



25- 



time 
for 

energy 
profit 20 



10- 



10 

energy ratio 



25 



Figure 7. The time for energy profit as a function of 
the energy ratio assuming a schedule of one 
station started each year. 



380 



36. 
The same building schedule could also be considered on the 
basis of fixing the input power to each type of reactor, 
then the output powers would vary with E . This ia 
demonstrated in figures 8 and 9, where it can be seen that 
all the curves are the same over the first 6 years but then 
diverge because of their different output powers. The 
variation between time for energy profit and E is exactly 
as in figure 7 only shifted down half a year. In practice 
alternative programmes are likely to be based on the same 
output power (although in a time of real fuel shortage the 
input power could be deliberately constrained) . Thus the 
first set of graphs are more pertinent. 

The results obtained from the numberical evaluations are 
consistent with the analytical solutions. The linear growth 
can be represented by 

where b is a constant. 



Then n c (T) = bT T 45 "f 4; 5» 

n^(T) = 5b t < 5 
n f (T) = b(T-6) T > 6 



/ / 



t*a 



To show a net power profit we require 

P o n f (t) > P i n c (T) 

„ E o v. n c (T) 
E F= E7 > n^TTl 

25 

This reduces to E ^ — , for a power profit. 

Hence to obtain a power profit in 10 years requires E R >6.25. 
(Compare this with figure 5) . To show a net energy profit we 
require E(T) to be greater than zero. This gives 

P o f n f (t) dt > pA n c (t)dt+P i l n^ 



(t)dt 



381 



37, 




JOMOd )8U 



382 



38. 




ABjeua jeu 



383 



39, 
Evaluating the integrals and re-arranging gives 

(50T-125) 



(T 2 -12T+36) 



Hence to show an energy profit in 10 years this tells us 
that you need an energy ratio greater than 23.44 (compare 
with Figure 6) . It should be emphasised that these 
results apply to any linear growth programme regardless 
of the number of reactors involved and the rate of linear 
growth - - the results are independent of the parameter 
b which defines the growth rate. 

Most projections of the growth in nuclear power capacity 
are not based on linear growth but on exponential growth. 
The historic growth in nuclear power in the U.K has been 
exponential with a doubling time of 4 years. Projections 
of the growth in nuclear capacity of the U.K (or Europe) 
by various authorities have the following doubling times. 

Doubling time Over period 

CEGB (30) 4.3 years 20 years 

UKAEA (31) 6 years 20 years 

UNIPEDE (33) 7 years 10 years 

( 10 years next 15 years ) 

OECD (23) 4 years 15 years 

ECC (32) 

3 years 12 years 

(23) 
The OECD report gives projections of world growth in 

nuclear capacity with an initial doubling time of 1.5 years 

falling to a doubling time of 7 years in 20 years time. With 

the emphasis of all these projections on fast exponential 

growth it is clearly this type of schedule that needs most 

examination. 



384 



40 



net energy 




Figure 10. The energy graphs for different energy ratios assuming 
a schedule with a doubling time of five years. The 
shaded blocks show when reactors aa?e started. 



385 



41 



1 , 



L- 



1 



L. 



l— 



r" 






' : x 

1 L.j...k^ 



[ 



I 



— A... 



r" 



d i 



jaMod ;au 



386 



42 



Figure 10 shows the energy graphs for a building programme 
with a 5 year doubling time. The initiation of reactors 
is shown by the shaded blocks and the total number of 
reactors started shown by the dashed curve. The curves 
for different energy ratios are seen to be more divergent 
than in the linear growth case. Of particular interest is 
the fact that the curve corresponding to E =5 disappears off 
the bottom of the page, indicating that with this rate of growth 
such reactors will never show an energy profit. If the 
exponential growth is subsequently slowed down or stopped 
altogether then there is a massive energy surplus. But 
whilst the growth continues the energy deficit continues to 
increase. This is an important dynamic effect since a reactor 
with E =5 would appear to be a "good investment". However 
like all investments it can be overdone. With an exponential 
growth (doubling every 5 years) this reactor requires you to 
continually invest more energy in the next set of reactors 
than you are getting back from those you have already built. 
This is confirmed by examining the power graphs, shown in 
figure 11. The effect of a discrete building programme is 
to make the power graphs vary erratically, but it clearly shows 
that for E equal to 5 the power never goes positive. 

Figure 12 shows the variation in 'time for energy profit' 
as a function of E deduced from Figure 10. It is necessary 
to assume a maximum time for repayment of 25 years since 
beyond this time reactor replacement is necessary. Thus 
for a growth with a doubling time of 5 years the minimum 
feasible energy ratio is about 7.0. 

This dynamic result can be confirmed by analytical methods 
similar to those used for linear growth earlier. The 
mathematics becomes more complex since it is necessary to 
construct exponential functions which are consistent with 
the physical situation (for example we require n (o)=0 
and n (l)=n,(6)). An approximate analysis can be done if 
the cumulative number of reactors started is represented 



387 



43 



by (e at -l). (Hence at t=0 I have not started any). This 
means that 

n. (t) = ae at 
4<t) =e a < t " 6 >-l 
n c (t) = e at (l-e" 5a ) 

This gives a criteria for profitability, in terms of the 
energy ratio, E R , 

where T is the time to profitability. The parameter a is 



a = Ln{2) 

so if t D = 5 years, a=0.139 etc. 

The inequality involving E n , T and a can be used to define 

_c — — 
a minimum energy ratio (E (min) ) which is needed to obtain 

an energy profit in a time T years for a growth rate defined 

by a. 

■»*■*> ■ ;;i»r„ } 

Putting a = 0.139 in this expression gives the equation of 
the curve plotted in Figure 12. The table below gives the 
values of E (min) for different growth rates and different 
times for energy profitability. Since the station lifetime 
is 25 years the E (min) for T=25y is an absolute minimum. 



This is a critereon for energy prfotability in the steady 
state and does not take into account the energy invested 
in the first 6 years of the programme. For a more thorough 
analysis see treatment by Price (34). 



388 



44 



Table 14 



Doubling 
time 

fc D 


Growth 

constant 

a 


E R (min) 


T = lOy 


T=15y 


T=25y 


10 

7 

5 

4 

3 
1.5 


0.069 
0.099 
0.139 
0.173 
0.231 
0.462 


9.15 
10.81 
13.52 
16.37 
22.7 
85.49 


4.77 

5.99 

8.08 

10.35 

15.65 

73.16 


3.02 
4.17 
6.21 
8.49 
13.87 
72.07 



At the beginning of this paper it was pointed out that there 
were reasons to question the amount of fuels being invested 
to make future supplies available. If we are entering a 
period of fuel shortages or constrained fuel supplies then 
the less the requirements of fuel investment the better. 
It was pointed out earlier that our current fuel investment 
is 2.5% of total fuel supplies. It is possible to use the 
present analysis to estimate the future fraction of fuel 
investment. The ratio of energy invested to output in any 
one year of a reactor programme is simply 



Investment ratio 



P o n f (t) 



Provided the programme has been underway for sometime this 
ratio is independent of time. (This assumes that the 
number of reactors finished is much greater than 1 so that 
exp (a(t-6) ) >>> 1) . Substituting for n (t) and n f (t) gives 

Investment ratio 



5(l-e 



389 



Table 15 Percentage of output which has to be 
invested in nuclear stations 
(equilibrium case) 



Doubling time 

fc D 

10 
7 
5 



3 
1.5 



45 





Energy Ratio 






< E R > 




5 


10 


15 


44% 


22% 


15% 


70 


35 


23 


115 


57 


38 


162 


81 


54 


274 


137 


91 


1140 


720 


480 



390 



46 



This gives the investment percentages listed in Table 15. 
Where the investment percentage is over 100% then this 
indicates a building schedule which is a loser all the way. 
(For instance with a doubling time of 5 years an energy 
ratio of 5 is a loser as shown by figure 10,11 and 12 and 
Table 15; . Even for the profitable programmes the proportion 
of output going to investment is substantial. A popular 
projection for the fuel supply pattern in the U.K by 1990 
is for nuclear power to supply over 10% of the total fuel. 
If this were the case then the energy investment in nuclear 
power as a fraction of total fuel supplies would be a 
tenth of the values given in Table 14. Since the 
projection requires a growth rate of 4% for 20 years this 
shows that the rate of fuel investment in nuclear power will 
be between 16% and 5% of total UK fuel supplies. If ' 
the other 90% of UK fuels are also derived from capital 
intensive sources, such as oil from deep oceans, gas from 
1 coalpleses then the total fuel investment in the UK could 
rise to 20-40% of the primary fuel input. This would have 
serious consequences for the quantity of fuel supplies 
available to final consumers. Worse still if in order to 
increase the supplies to final consumers (i.e. to try to 
compensate for the fuel lost to investment) the rate of 
growth were increased then the energy investment would rise 
still further. 



A more thorough analysis of the dynamics of the exponential 

(34) 
growth schedule is given by Price" 



In many situations the discrete growth of reactors does not 
conform to an exact analytic function. Under these 
circumstances it is necessary to use the simple computer 
programme based on discrete building schedules. For 
example figure 13 shows the building programme proposed by 
the CEGB evaluated for a range of energy ratios. (The 
particular numbers corresponded to energy ratios for reactors 
in the initial report. The new energy ratios given in Table 
12 of this Report can be interpolated on this figure) . The 



391 

47 



significant point about this building schedule is that 
it only begins to show an energy profit after the peak 
of construction work. If the growth in nuclear capacity 
projected by the CEGB is to be sustained another building 
programme would have to be instigated about year 10 and 
this would further delay the time of energy profitability, 



392 



= 1 'S 

< < 5 e> 
o 5 a. </> 






>> 




£t m 




M -H 




-D 


-h -a 


M-i <U D) 


» 0) 


D^+J >-( 


CN +J 


W C fO 


- ra 


x; -h o +j 




a co o u 


■> 3 


m 3 > c 


CN CJ . 


M 'O <u 


~rH 2 


D to nj v* 


~< ro s 





- o 


b-P rH O 


CN O 


- >.o 


U ft O 


ro tr>o 


t) ht $ 


« )-i -i 


C (DO 


■v <D 


CD >iXI C 


- c cn 


Cn o a) 


m w -h 


<D M M 3 




XI <U Cd — x: 


En C CT> d) 


- • O 


CD C W 


ro kf> i0 


•H ^ 


»ro Q) 


f>|.p -O • 


m 


3 -h o 


»U-I O^ 


CN O C 


0) 


U 3 W 




M 


CU XI U 


CNr-t e 

- co 3 


3 


U-l 


tr **-i CD CD 


N 4J (/I 


•H 


•■h x: xj 


» O W 


[n 


>a -p -u cn +j ro 



48 



L 



" • 






i 








i : 

i : 
I-V-. 






) 

i 
1 






r 






jj 




r 
















i ; 




> 
, i 








i 
-J 





















(<MA) A6j»u»}9u 



■►I 



393 



49 



4. 3 Future Plans 

As indicated in the preface this is only an interim report 
of an ongoing research programme. Our principle conclusions 
from the investigations to date are the areas which require 
further study. If nuclear power stations are built with 
the purpose of adding to the total fuel supplies of a society 
then the energy analysis of reactors and building schedules 
should provide useful policy information. In practice 
the situation is more complex because, for example, the UK 
does not mine its own uranium and thus does not expend this 
energy. Whether the political dependence on uranium exporters 
is better or worse than a dependence on oil exporters is an 
important issue which energy analysis can do nothing to 
resolve. What energy analysis can do is to give policy 
makers a good idea of the energy implications of future 
policy. For instance it would appear from this analysis 
that the crash nuclear programme started by France may 
require them to substantially increase their imports of oil 
for 10-15 years. We suspect that the French policy makers 
had no idea that this was even a possibility. 

Another major conclusion of this type of analysis is to bring 
into question the'quasi -infinite energy resources" represented 
by the minute concentrations of uranium in natural rocks 
and sea-water. From an energy analysis point of view there 
is a cut-off grade of uranium which requires more energy 
for its conversion to a fuel rod than can be obtained from 
the fuel rod itself. We believe that this report provides 
substantial evidence for this view and puts a serious 
constraint on the quantity of uranium available as a real 
energy resource. (In this respect very low grade uranium 
resources are akin to the oil left in the bottom of a well 
at the point where it takes more energy to extract the oil 
than the extracted oil contains.) On the scanty evidence 
which we have to date it appears to us that the minimum cut 
off grade is of the order 100 ppm U-Op. (0.01%) . This is 



394 



50 



based on the fact that to achieve a growth in nuclear 
capacity of the order 10% per annum (t n ?6 7 yrs) requires 
a minimum energy ratio of about 5. For Chattanooga shales 
(0.00 7%) only CANDU has an adequate energy ratio, all the 
other reactors would loose energy using this ore whilst the 
growth was 10% per annum or more. 

This type of reasoning raises more questions than it answers. 
For example we do not know how characteristic Chattanooga 
shales are as a low grade source of uranium. Nor do we know 
how soon this type of ore grade will have to be exploited. 
According to TNPG this grade of ore could be needed by 
1995 which is within the lifetime of reactors being 
built today! There are a number of conservation strategies 
which could change the details of the picture. For example 
by using much lower tails assay the present uranium reserves 
could be made to last longer. The burner reactors could use 
the plutonium they produce as a fuel again decreasing their 
uranium needs. But before anyone can begin to assess whether 
these are sensible possibilities or not the following information 
is absolutely essential. 

(i) data on the energy required to mine uranium at 
many different ore grades. (Hence construct a 
graph such as that shown in figure 14.1). 

(ii) data on the variation in reactor energy ratio 

as a function of ore grade. We have found that 
at high ore grades a tails assay of 0.2 5% gives 
the best E but at low ore grade 0.2% tails 
assay gives the best E R . Where is the 
changeover? (Hence construct the details of 
figure 14.2) . 

(iii) Refine the accuracy of the energy analysis by 

doing process analyses on the principle reactor 
components and any special materials needed. 



395 



i 

k 

r 

V 



Urc^o^ or* gf-ewk- (% U,0*) 






av^y )AdV- bfi 






51 



0kcc fu^ctw^ lb* rVso^« ^ ft* cur«tf. 
cjvmJ- arc rtu l*Afna»ha«x •£ a *H i *„i - 



3 

3 



P*0fc (o.tf% fault) ^ 




1 1 r» 

Ufa«\lu«M Ore $r»de (% UjC^) 







I 



•JH 



TKtS f«43 «M ttwi fffc*v «L*K^ atoto am 



396 



52 



This data could then be combined with the curves shown in 
figures 14.3 and 14.4 to evaluate a wide range of possible 
nuclear strategies based on different reactors or/and 
different rates of growth. This is the ultimate goal of 
our research programme. 

5. Conclusions 

"There's no use trying", she said, 
"one can't believe impossible things." 
"I daresay you haven't had much practice" 
said the Queen. "When I was your age, I 
always did it for half-an-hour a day. Why, 
sometimes I've believed as many as six 
impossible things before breakfast". 

(L. Carroll) 

The above quotation is a succinct summary of the initial 
reception of our report by many economists and engineers. 
Energy analysis provides a new way of looking at the world 
and when its conclusions contradict commonsense there is 
a suspicion that the method is crazy and producing impossible 
answers. However the conclusions of this analysis are only 
"impossible" if foolishly interpreted. The final section 
of this report is an attempt to clear away many of these 
misinterpretations. 



Our basic conclusions are 



(i) nuclear technology is a good way of generating 

electricity (since it requires less than 4 kWh of 
coal energy to produce lkWh of electricity) . 

(ii) only fairly rich uranium ores (concentration > 0.01%) 
should be incorporated in estimates of total energy 
resources. 



397 



(iii) that in a period of growth in the number of reactors 
on energy ratio greater than one for a single reactor 
does not guarantee an overall energy profit for the 
building programme. The rate of growth of the building 
programme sets a lower bound on the energy ratio 
required to produce an energy profit in a specified 
time. 

(iv) that in the future the most important parameter 
determining the profitability of nuclear power 
stations is the energy involved in mining and milling 
uranium. (For low grade ores enrichment energy is 
relatively unimportant) . 

(v) that even with an energy profitable nuclear programme 
the proportion of annual fuel supplies required as 
fuel investment may rise to 20-40% in the next 20 
years. This has serious implications for the supplies 
of fuels to domestic and industrial consumers. 



53 



398 



54 

6. Comments and answers 

There have been many different reactions to our initial report. Amongst 
these reactions there are a number of recurring questions which we think 
are worth answering here. 

I. "If this analysis of nuclear power programmes is correct in showing 
that they can be net consumers of energy how is it that they appear 
economically attractive? (Why does energy analysis give a different 
answer to that from economic analysis?)" 

This question, or variations on it, is the most common reaction to the 
analysis. There are three points which need to be made in reconciling 
the apparent contradiction between energy and economic analysis. 

a) The economics of nuclear power are not as clearly favourable as some 
pundits would make out. Most economic analyses of nuclear power have 
been done in order to promote nuclear technology. This bias is 
reflected in omitting to take into account items such as the capital 
costs of enrichment plants and the increase in uranium costs due to 
increases in conventional fuel costs. It is, however, true to say 
that when costed properly nuclear power does appear preferable to 
fossil-fired power stations. 

b) The comparison made in economic analyses of nuclear power is between 
nuclear generated electricity and electricity generated from fossil 
fuels. However everyone knows that fossil fuelled power stations are 
net consumers of energy; they consume between 3 and 4- KWh of fossil 
fuel for every 1 KWh of electricity sent out. So if a nuclear station 
consumed say 2 KWh of fossil fuel for every KWh of electricity sent 
out it would appear, in energy and economic terms, a preferable 
technology for generating electricity. However under these conditions 
you would count neither fossil-fired stations nor nuclear stations as 
methods of increasing the world's fuel resources. (To show, by 
economics, whether nuclear power was a net producer of energy would 
require a demonstration that the price per therm from a nuclear power 
station was less than the price per therm from a coal mine or oil 
well). 

c) The final point concerns the overall objectives of energy analysis. 
It is not a system of allocating resources, it is not a replacement 
for economics. All that energy analysis can do is to tell you the 
energy implications of policies. In contrast economic analysis tells 



399 



55 

you the financial implications of policies. There is thus no 
contradiction between the methods of analysis, they complement each 
other. The mistake implicit in the question was the assumption that 
a fuel industry which makes a profit must be a net producer of energy. 
The CEGB is a good example of why this assumption is wrong. Energy 
is just one of the resources considered in economics and can only be 
considered the most important if the policy being analysed specifically 
states an energy objective as being the most important. 

Thus, in summary, the answer to the question is that nuclear power shows 
a profit when compared to an industry which is a prolific net consumer 
of energy but this does not imply that nuclear power is a net producer of 
energy. 

P. "If your analysis is correct then what should be done?" 

The simple answer to this is nothing. Our analysis shows that you may not 
increase fuel supplies by building nuclear reactors - so do not build 
them. To show that a policy is mistaken does not oblige us to have an 
alternative policy. 

3. "Your anlysis has confused 'energy' with 'welfare' and it is not 
obvious that this type of analysis is at all relevant to nuclear 
policy". 

This is not a constructed comment, it is an accurate summary of comments 
received from the Department of Energy. The reply to the comment was 
that the report nowhere confused welfare with energy and that the only 
justification for looking at the energy implications of nuclear power 
stations was because it was claimed (by their proponents) that they 
increase the supply of energy. If nuclear power stations are constructed 
for other reasons then energy analysis may well be irrelevant to future 
policy formulation. 

4. "You have only demonstrated that nuclear systems are net consumers of 
energy under ridiculously high rates of growth. If you were to use a 
more sensible growth rate then your conclusions would be different". 

The official estimates of nuclear power requirement, by the CEGB, UKAEA 
andoECD all conform to a growth rate with a doubling time of 4-J years for 
the next 25 years. This is not because this is the rate of growth of 
total energy demand. It is because nuclear power is being put forward as 



400 



56 



a source able to take over the role of oil and other fossil fuels. We 
agree that these rates of growth are ridiculously high however they are 
the historical growth rate (31) and the official projected growth rate (29) 
(If you want a new technology to take over as an old technology declines 
then the rate of growth in the changeover period has to be very fast. 
This is illustrated in the figure below) 







*Toh«.l tfv^vgcj 


f 






I 

< 


—~^\ 


/* 


$ 


foss t \ 


/\ 




Fu*(s 


J 

:> 



H*« 



r j. "Even if you have got the correct growth law no one is pretending 
that this will go on indefinitely since breeder reactors will take 
over the growth within 15-20 years". 



We do not pretend that the growth will continue indefinitely. We are 
essentially analysing a transient problem, the problem associated with 
the period while the nuclear burner programme is expanding rapidly. We 
have only shown that nuclear reactors (using present uranium ore grades) 
are net consumers of energy during a growth period. Their viability in 
the future will depend upon uranium ore grade and any technical changes 
in nuclear technology. 

If breeder reactors are to take over a large part of energy demand then 
you will have to continue to increase the number of burner reactors. 
This is because the 'breeding time' for breeder reactors is 20 years, so 
on their own they are constrained to a doubling time of 20 years. To 
build up a stock of breeders at a faster rate than this means that some 



401 



57 

source of plutonium other than breeder reactors has to "be used. The 
only other source of plutonium is a burner nuclear reactor. So to make 
a fast growth in breeder reactors possible it is necessary to first have 
a f'.-ir.t growth of burner reactors and for the programmes to run concurrently 
lor a very long time (50-100 years). 

There have also been a number of more detailed comments and questions 
which are worth clearing up at this point. 

6. "Why is the initial fuel counted as an investment cost?" 

Because you cannot start a nuclear power station until the entire core 
has been assembled. This is thus a true 'investment' since all the 
energy (and money) for the core has to be paid out before any energy can 
be obtained from the reactor. The initial core actually represents about 
an o i i';hth of all the fuel required by the reactor. For a coal-fired 
station this would be equivalent to a situation where 14 million tons of 
coal had to be mined and delivered to a 1000 MW station before it could 
be started up. 

7. "Since the core is intact at the end of the reactor life time this 
should be included in the total output of the reactor. 

If a reactor core is still intact when the reactor is decommissioned then 
it could change the energy ratios for second generation reactors. It 
has absolutely no effect on the transient problem analysed in the paper 
(apart from anything else in a growth period the number of reactors being 
started up is always much larger than the number being closed down). 
It is also doubtful whether anyone would continue to refuel a nuclear 
reactor right up to the point of shut-down. When a reactor is 
decommissioned it is more likely that a fraction of its core will be 
useful fuel for future reactors. 

8. "You have not given any credit for the plutonium produced". 
This is because :- 

a) If you count the energy value of the Plutonium as a credit then a dtfbit 
must be conbedfor the energy value of the uranium as a 
material. In both cases the raw material is of no energy value to 
anyone until it has been converted into a useful form. We are 
concerned with conversion costs and not the potential energy of the 



402 



58 



raw material. 

In terms of real energy flow into an economy the plutonium is 

irrelevant. We have shown that a growth programme in burner reactors 

consumes more energy than it produces. It is not much use offering 

rods of plutonium to people who want electricity. So really 

plutonium is irrelevant to the transient problem we are concerned 

with. 

b) The plutonium is not normally considered as a fuel for burner 
reactors (though it could be). Any credit given to burner reactors 
lor producing plutonium would be a debit set against breeder reactors. 

c) Wo do not have any data on either the energy to extract plutonium 
Crom spent fuel rods nor the energy which could be derived from the 

plutonium. 

9. "You have not included any energy used in waste handling or storage". 

This is because we have no data on these operations. Any energy 
expenditure on these operations would only make the situation worse than 
wc have painted. In the long term these energy expenditures may be 
important but we doubt whether they would significantly alter the nature 
of Lhc transient problem analysed here. 

The most important waste material is plutonium with a half-life of 24,000 
years. Although plutonium is recovered from spent fuel rods 100% recovery 
is not economic (probably not energy profitable either). Thus waste from 
spent rods may have to be stored for 240,000 years (10 half-lives). Now 
if a power station produces 1000 NW for 25 years but leaves waste materials 
which require a power input for maintenance of 100 KW for 250,000 years 
then the net energy output will be zero. A power input of 100 KW is 
equivalent to an annual consumption of 90 tons of steel or 25 tons of 
rolled stainless steel. Hopefully designers of waste-handling will be 
able to keep the average power consumption below 100 KW as well as develop 
an infallible system capable of surviving for 250,000 years! 

10. "You have counted electricity consumed in enriching fuel at its 
thermal rate based on conventional fossil fuelled stations. However 
as each 1000 MW reactor is built the average efficiency of generating 
electricity will get better and better, thus making the energy costs 
of nuclear reactors smaller and smaller". 



403 



59 - 

This is not a valid comment because whilst the nuclear power programme is 
still running at a net energy loss the efficiency of generating electricity 
is actually going down. Thus by maintaining one efficiency throughout the 
growth period we have been overgenerous. If ever a nuclear programme moves 
into net energy profit then it would be correct to use an improving 
efficiency for generating electricity. 

11. "You have only considered gas-diffusion enrichment of uraniam. If 

you had considered gas centrifuge enrichment, which requires a tenth 
of the power input, your conclusions would be changed". 

Wc are aware of the technology of centrifuge enrichment but do not have any 
operating data to confirm the estimated power requirements. Even if the 
centrifuge has a tenth of the operating power the fact that it is more 
capital intensive may not affect our conslusions. At the moment the 
enrichment process is the most energy intensive step in uranium processing. 
The capital costs of gas-diffusion enrichment contribute 10% to the total 
energy cost of enrichment. The centrifuge process has been estimated to 
be more capital intensive than gas diffusion. Further more the other 
operating costs of centrifuge plants are much higher than for gas-diffusion. 
So the energy cost by centrifuge enrichment is likely to be at least 50% 
of the energy cost we have used for enrichment. This would improve the 
energy ratio for any one reactor. However the shift of energy cost from 
operating costs to capital costs would make the growth problem worse. 
Thus until we have better data we cannot be sure whether the centrifuge 
process represents a net improvement in the energy flows or not. 

1?. "You have included uranium mining in your energy analysis even though 
the UK does not mine any uranium. Surely this makes nuclear power 
very attractive as a net energy source for the UK". 

The energy cost of mining uranium is not a significant item when based on 
preset ' ore grades. Thus taking this term out of the energy accounts 
would only improve the energy ratios by about 0.2 which is within our 
error estimates. The important energy cost items, fuel enrichment and 
station construction, are bourne by the UK. 

In the future, as the grade of uranium ore declines, the energy cost of 
mining will become more and more significant until, at grades less than 
0.01%, the mining energy dominates. At this point you have to ask whether 
it is reasonable to expect uranium exporters to subsidize your fuel 
production system. This is essentially a political question, energy 



404 



60. 



analysis only shows you the fuel implications of the situation. Basically 
you will have to decide whether a dependance on uranium exporters is more 
or less preferable to a dependance on OPEC1 



References 



405 



61 



"Energy for the future" 



The Institute of Fuel: 18 Devonshire 
Street, London W.l. 1973 



2. Chapman, P.F. 



Leach G. and Slesser, M. 

Energy Policy 2 (3). Sept. 1974 



?. Chapman, P.F. 

4. Searby, P.J. 

!). CKGB evidence to. 



6. Chapman, P.F. and Mortimer N. 



7. Wright, D.J. 



Energy Policy 2 (4) (in press: Dec. 1974) 

Atom 17_8 (Aug. 1971) p. 185 

"Select comm. on Sci and Technology; 
The choice of reactor system" 
(15BN 10 276574 x) HMSO 
Appendix 6 p. 192 



"Energy analysis of the Census of 

production 1968" 

O.U. Research Report ERG 006 (in 

preparation) 

"Calculating energy requirements of 
commodities form the input/output table" 
SARU: Dept. of Environment London S.W.I. 



Everett, F.D. 



9. Bieniewski, C..L. et.al 



10. 



11, 



12. 



Clegg, J.W. and Foley, D.D. 



Jiarnaby, C.F. 



"Mining Practices at 4 Uranium Mines" 
U.S. B.O.M. Inf. Circu. 8151 

"Availability of Uranium at Various 
Prices from Resources in the U.S." 
(U.S. B.O.M. Inf. Circ. 8501. (1971) 

"Uranium Ore Processing" 
Add i son-Wesley 1958 

"Fuel Cycles for Electrical Power 
Generation" Teknekron Inc. California. 
Jan. 1975 

Sci. Journal. Aug. 1969 SA (2) p. 54 



406 



62 



13. Roberts, J.T. Int. At. En. Bull. lj? (5) Oct. 1973 p. 14 

14. "Directory of Power Reactors" 

l'>. Moore, J. Bradly, N., Rowlands, I.T. 

Atom 12£ (Jan. 1973) p. 7 

16. "Deuterium" in Kirk-Ohmer 

"Encyclopaedia of Chemical Technology" 
McGraw Hill 

17. McTighe, E.P. "The development of the UK nuclear 

power industry" Atom 170 Dec. 1970 p. 242 



18. Patterson, W. and Lovins, A.B. 



19. Hcf. 5 p. 193 



Evidence to Select committee on Science 
and Technology (ref.5) p. 152 



20. Details in "Report on Census of Production 1968" vol. 153. 
(Table 5) : see also Chapman P.P. in "Conservation of Materials" 
proc. of conference. Harwell. 26-27 March 1974 p. 125 

21. Chapman, P.F. "Energy Accounting" in proc. conf. on 

"R and D routes to effective energy 
utilisation" NEL Sept. 1974 

22. Data supplied in letter to G. Leach from the Manager; Heavy Water 
Plant Technology; At En. Canada Ltd. 

?3. 'Uranium; resources production and demand 1 OECD. Aug. 1973 

24. Avery D.G. and Kehoe, R.B. Atom. 164 June 1970 p. 120 

25- Benedict, M and Pigford, T.M. 'Nuclear Chemical Engineering' 

Mc.Graw Hill 1957 

26. Index of Nuclear Reactors. 'Nuclear Eng. Intl' April 1972 

27. Data from Gentilly Station Nuclear Eng. Intl (supplied by UKAEA) 

28. From estimates given by UKAEA and TNPG. 

29. Sir John Hill. (Chairman UKAEA) in Atom 180 Oct. 1971 P-231 



407 



63 

30. Based on graphs given in ref. 5- « The nuclear growth is much faster 
than total electricity growth since it is replacing oil and coal. 

31. Bainbridge G.R. and Beveridge C. Atom. 133 Sept 1969 p. 248 

3?. Based on energy policy plans reported Guardian 29. 11. 74-. To get 

to 45% nuclear by 1985 requires 13 fold increase in nuclear capacity 
in 11 years. (This assumes total electricity demand doubles in this 
period. 

33. Prospects for Long Term Dev. of F.B.R. in the E.C. 

UNIPEDE March 1974 

Dynamic energy analysis of nculear 
reactors. Dec. 1973 F.O.E. (to be 
published) . 
35. Vaughan, R.D. The need for F.B.R. ■ -in Europe 

Dec. 1974 (to be published). 



408 



yr 



^T75' y £ ' 291.7 



Drill bits and steel 



780 x 1CT .^_. x 1 x 220 /KWht/ x 1 

yr' ITS ( £ ' 291.7 



336 x 10 3 ,i , x 1 x 50 ,KWht * 3 ' x 1 

( f? ) 23 ( T- > 291T7 

Operating supplies 

147.8 x 10 3 .£_. x 1 x 80 , KWht * ' x 1 

l yr 23 { £ ' 291.7 

Mine plant costs 

(See footnote 5 and Table D-2) 



SUMMARY OK CALCULATIONS OF MINING CHATANOOGA SHALES 

(PFC 3/75) 

Mining 

Table D-ll (p 88 attached photocopy) gives mining data. 
Annual production U„0_ = 653,400 lbs/yr = 291.7 tonnes/year. 

3 o 
(See footnote 2) Mine life 20 years; thus total production 5736 +ALO ft 

Electricity KWht/ton U f 

7120 KW x 7200 -hr. x 4 /KWhtv ' x 1 ( year > = 703 x 10 3 
yr' l KWhe' 291.7 tons' 

Water 

1.8 x 10 3 . gall , x 7200 .hr. x 10~ 2 , KWht * 2 ' x 1 = 0.4 x 10 3 
1 hr ' V gall' 291.7 

Tunnel machine supplies 

1.296 x 10 ,&_. x 1 ,£. x 50 /JWht/ 3 ' x 1 = 79.3 x 10 3 

( 7? ) 23 ( ? ) ( "£- ) 29T7 

Explosives 

660 x 10 3 ,$ . x . 1 v x 120 ,KVht* x 1 = 97.2 x 10 3 



Diesel fuel 

(1) 3 

2200 /gall* x 300 , day , x 54.5 , KWht . v ' x _1 = 123.3 x 10 J 

May ' V ' gall' 291.7 

Roof bolts 

218.7 x 10 3 ( £ ) x 1 x 80 /KWhtv x 1 = 21.4 x 10 3 

>ear' "O ( £ ' 291.7 

Mine maintenance materials 

(3) 



* - 1 x 50 / KWht . 3 x 1 " * - ' " 3 

"O" l £ ' 573? 



21.944 x 10 £ x __1_ x 50 . KVht j ' x 1 68.3 x 10 



TOTAL FOR MINING 1338.0 x 10 3 



409 



Milling 

(Data from Table D-12) 

Electricity KWht/ton U f 



(1) 3 

13870 KW x 7920 ( hr> x 4 , KWht } x 1 = 1506 x 10 J 

yr l KWhe' 291.7 

Water 

( 9ail ^ X l^f£AJ 

yr' 'gall' 291.7 
Coal (for steam raising ) 

400 . tons , x 0.89 ( te > x 330 . days , x 8334 , KWht . (1) x 
l day ' K VS ton' yr ' Hon ' 

Rods for mill 



528 x 10 3 - lbs , x 1 . ton , x 40,000 .KWht. x 1 

l yr 2240 l lbs' Hon ' 291.7 

Chemicals 

(Breakdown into types given in table D-14) 

N-B-A 

(3) x 1 



56.8 x 10 3 .*> x 1 x 450 ,KWht ' 
( f7 } 2^ ( -£- ) 

Kerosine (fuel oil) 

95700 . gall , x 1 . ton . x 15279 f KWht . (1) x 1 

yr ' 2BT gall' Hon ' 291.7 



Sodium carbonate 

2626800 ,1b. x 1 . ton , x 7000 /KWht* ' x 1 

yr 22^0 Ubs Hon 291.7 

Anhydrous NH 



102.3,tons. x 0.89 ( te . x 15,000, KWht j ' 



yr US ton tonne 291.7 



'.70'il5 . tons . 0.89 ( te . x 400 . KWht . ( * x 1__ 

yr X K VS ton' Hon ' 291.7 

(Note if done on financial basis, 

16.229 x lO 6 £_ x 1 x 19^.5 KWht x 1 = 3864 x 10 3 ) 
yr "O" £ 291.7 



68-391 O - 76 - 27 



410 



278" l £ ; 5736" 



llocculent 

92.4 x 10 3 £_ x 1 x 19**. 5 , KWht . (3) x 1 

yr "O" l £ ' 291.7 

Mill materials 

168 x 10 3 ,g_. x 1 x 50 /KJJhtv x 1 

yr' 278" l £ ' 291.7 

Plant and capital equipment 

(Based on 50.9 x 10 $ for plant and 11.83 x 10 % 
for utilities as shown in Table D.4) 

Plant 

50.991 x 10 % x 1 x 50 , KWht > 3 x 1 

"278" K £ ' 5736 ♦ U 3 0g 

Facilities and utilities 

11.829 x 10 6 # x _1_ x 35 ( KWht ^ 3 * x _1 25.8 x 10 3 



TOTAL FOR MILLING = 5773.4 x 10 3 



GRAND TOTAL 7111. 4 x 10 3 KWht/ton U„0 o 
3 o 



( Note this assumes 70% efficiency taken into account in data. 

If only 70% of this U„0 o available then cost rises to 10159 x 10 3 KWht/ton 
3 o 



411 



1. Baaed on values in attached paper front Energy Policy. See especially 
Table 3. 

2. See ERG 001 for derivation of this. 

3. Based on analysis of Census of Production 1968. Note that we also 

use 1968 value of $ to be consistent. The results from our analysis are 
similar to those of Wright in D.O.E. based on I/O table and to values deduced 
for USA by Herendeen (CAC Illinois) using the US I/O table. Important 
numbers are: 

Mining machinery 50 KWht/£ value 

Plant steelwork 35 KWht/£ 

Explosives 120 KVht/£ 

Steel 220 KVht/£ (see Energy Policy March 1975) 

Nuts and bolts 80 KWht/£ 

Inorganic chemicals 194.5 

Organic chemicals 454.7 

4. This is a rough average of industrial 'materials'. See also Energy Policy 
March 1975. 

5. This is based on fact that mill rods have to be hardened steel with 
estimated energy cost 3x higher than average (which is 13200 KVIht/ton - 
see Energy Policy March 1975)* 

6. Deduced from Data on process inputs in Shreve, R.N. 'Chemical Process 
Industries' 3rd Edition McGraw Hill. 

7. Based on analysis by Leach and Slesser. They give a value of 13000 KWht/ton 
of liquid ammonia and say that anhydrous NH is significantly more energy 
intensive. 

8. This is the smallest energy cost/ton of any industrial material. It is 

the same as the energy cost of steel scrap when delivered to a steel worker. 
Since scrap and H0SO4 require zero energy for their production but involve 
energy costs in storage/capital plant/ transport etc this value seems 
reasonable. It is certainly more reasonable than that based on the 
financial calculation (which implies an energy cost of 2700 KWht/ton). 
To put a more accurate value on this input would need to know, for 
example, exact transport distance since if H2SO4 plant is 100 miles 
away then at 1 KWht/ton-mile the 200 mile round trip would contribute 
200 KWht/ton. Note that by chance the contribution from H2SO4 to the 
total energy cost of milling is 10#, so it is not a drastic problem unless 
I am underestimating by a factor of 2 or more. 



412 



DYNAMIC ENERGY ANALYSIS 
AND NUCLEAR POWER 

BY DR JOHN PRICE 



18 December 1974 

Friends of the Earth Ltd for Earth Resources Research Ltd 
9 Poland Street London W1V 3DG 



413 



NOTES ON THE PUBLISHER 

Earth Resources Research Ltd., founded in London in 1973, is an 
educational and research charity associated with Friends of the 
Earth Ltd. This in turn is the independent British member, founded in 1970, 
of a worldwide family of non-profit making lobbying groups dedicated to 
the conservation, restoration and rational use of the earth. FOE Ltd., 
with a full-time staff of 14 and with lOO-plus local groups from Cornwall 
to the Shetlands, actively pursues specific campaigns on (inter alia) 
the control of packaging, the protection of endangered species, National 
Park and land-use policy, and resource and energy policy (particularly 
energy strategy, North Sea oil, nuclear policy, and the mining and con- 
servation of metals). FOE Ltd. is one of ten national members of the 
"umbrella" organisation FOE International, accorded Non-Governmental 
Organisation status by the United Nations and specialized agencies. 



Earth Resources Research Ltd., 9 Poland Street, London W1V 3DG 
(01-434-1684) 



Printed by: Press-on Printers (Herne Hill) Ltd., 6-7 Oak Place, East HiH, SW18 2QB 



414 

DYNAMIC ENERGY ANALYSIS AND NUCLEAR POWER 

BY 

JOHN H PRICE 

18 DECEMBER 1974 



An initial inquiry (intended for the layman) 
into how the net energy balance of exponential 
programmes of energy conversion facilities 
varies in time; what are the energy inputs and 
outputs of commercial nuclear reactors, both 
singly and in such programmes; what are the 
possible errors and omissions in this analysis; 
and what are the policy and research implica- 
tions of the results. 



CONTENTS 

1. Introduction: why energy analysis matters, 1 

2. Energy evaluation of energy conversion processes, 2 

2.1. Energy profitability of a single energy 
conversion plant, 2 

2.2. Energy profitability of a programme of 
plants: a simple example, 3 

2.3. Energy profitability and exponential growth, 5 

2.4. Summary, 12 

3. Static energy analysis of nuclear reactors, 12 

3.1. Energy investment in capital plant, 12 

3.2. Special investment inputs and some corresponding 
process inputs, 13 

3.2.1. Uranium fuel, 13 

3.2.2. Heavy water, 14 

3.2.3. Other inputs, 15 

3.3. Output of energy from a nuclear power station, 15 

4. Dynamic energy analysis of nuclear power programmes, 16 

4.1. Energy profitability of exponential nuclear 
power programmes, 16 

4.2. Critique of methods and data, 17 





4.2.1. 


Investment inputs, 17 




4.2.2. 


Process inputs, 18 




4.2.3. 


Other inputs, 18 




4.2.4. 


Outputs, 19 




4.2.5. 


General methodology, 19 




4.2.6. 


Summary, 20 


5o 


Some questions 


and answers, 20 


6, 


Conclusions and policy implications, 23 



Appendix: Exponential growth: an analytic treatment of 
energy conversion programmes designed to 
satisfy exponential demand, 26 

References, 28 

Author, 30 

Acknowl edg erne nts , 30 



Copyright © 1974 by John H Price. This preprint is 
distributed by Friends of the Earth Ltd (for Earth 
Resources Research Ltd) , 9 Poland St, London W1V 3DG, 
from whom additional copies are available for £1. each 
(postpaid in UK only) . 



415 



"Nuclear power generation must take over a large part of the increase in 
energy demand by 1980." 

— The Institute of Fuel 



"'There's no use trying', she said, 'one can't believe impossible things. 

'I daresay you haven 't had much practice ' , said the Queen. 'When I was 
your age, I always did it for half-an-hour a day. Why, sometimes I've 
believed as many as six impossible things before breakfast. '" 

— Lewis Carroll 



"Decrease does not under all circumstances mean something bad. Increase 
and decrease come in their own time. What matters here is to understand 
the time and not to try to cover up poverty with empty pretense. If a 
time of scanty resources brings out an inner truth, one must not feel 
ashamed of simplicity. For simplicity is then the very thing needed to 
provide inner strength for further undertakings. Indeed, there need be 
no concern if the outward beauty of the civilization... should have to 
suffer because of simplicity. One must draw on the strength of the inner 
attitude to compensate for what is lacking in externals; then the power 
of the content makes up for the simplicity of form. ...Even with slender 
means, the sentiment of the heart can be expressed. " 



I Ching 



416 



DYNAMIC ENERGY ANALYSIS AND NUCLEAR POWER 

JOHN H PRICE 
18 DECEMBER 197^ 



1, INTRODUCTION: WHY ENERGY ANALYSIS MATTERS 

Since about 1971, the staff of Friends of the Earth have been studying problems of energy policy. A team of 
experts, "Energy 2000+" , was assembled about a year ago to examine the strategic options available to the UK. One 
member of this team, Dr Peter Chapman of the Open University's Energy Research Group, has for several years been 
working with colleagues to determine how much energy is needed to produce various goods and services, especially 
energy itself: mining raw materials, putting up buildings, and preparing fuel all consume energy which must in 
some way be offset as a debit against the output energy of the plant in calculating its net output of useful work 
for society. Of course, these energy inputs can be, and often are, supplied by some energy source other than the 
one being analyzed; but they still come out of society's total energy budget and must be debited somewhere, for in 
effect, the new energy source is on balance supplying less energy to society than it was supposed to do. We 
must therefore find out how much energy our energy technologies consume at what times; we can then calculate the 
energy versions of a cash-flow account and of a prof it-and- loss account. 

This report outlines the methods used to perform this energy analysis (§2), then analyzes six types of 

nuclear reactor systems both as individual power stations (§3) and in programmes (§4). The word "programme" 

denotes a sequence of reactors built according to some pattern, and dynamic energy analysis is concerned with 

the profitability (in terms of energy) of the entire programme as a function of time. 

Originally it was intended that the findings of this analysis should appear in a forthcoming paper on the 

2 
viability of the nuclear energy option for the UK. That paper will complement a recent assessment of certain 

nuclear policy issues (accidents, sabotage, waste management, security of strategic materials, etc) by consider- 
ing economies and diseconomies of scale, social and economic implications of centralization, and other broad 
problems. However, energy policy decisions are now being made in which the energy profitability of nuclear 
power systems and programmes has apparently not been considered at all. The implications of Dr Chapman's work 
are so important for these current decisions that it would be irresponsible not to make them known now. The 
author, Dr Chapman, Mr Amory Lovins, and several colleagues also plan to prepare over the next few months a 
technical version of this paper (more complete and rigorous than this outline for the layman) for publication 
in a scientific journal. 
have been incorporated i 
of programmes designed to satisfy exponentially increasing energy demand. 

Let us begin with some comments on the nature of the problem facing energy supply utilities. 

The energy and money needed to obtain utilities' raw materials (crude oil, coal, uranium, etc) are tending to 
increase as the higher-grade, more easily won reserves are depleted. For example, the capital and operating costs 
of extracting a barrel of oil from the North Sea are about ten times those for extraction from the Middle Eastern 
fields 3 . 

At the same time, the demand for energy, especially in the form of electricity, is officially predicted to 

increase rapidly, and the energy supply utilities claim to have to try to meet this projected demand. Figure 1 

shows the demand for electricity in England and Wales until the year 2000 as projected by the Central Electricity 

Generating Board. Figure 1 also shows how the CEGB plan to meet this demand: they expect nuclear power to take 

over the main burden of electricity supply as fossil fuels become less readily available. The combination of 

assumed rapid growth in demand together with declining availability of fossil fuels implies exceptionally rapid 

sustained growth (Figure 2) for nuclear power; and it is the burden of this paper that such conditions raise 

important and novel problems which dynamic energy analysis is well-suited to address. Unfortunately, the data 

available for such analysis are sometimes of dubious quality or have to be estimated indirectly from financial 

data. The author nonetheless believes that dynamic energy analysis lends much confidence to the conclusion that 

traditional views of nuclear power as an abundant source of energy in the future are a myth. 

4 
This myth finds implicit expression in a statement by Professor D C Leslie , who, after acknowledging that it 

is impossible to quantify the risks or consequences of accidents, remarks: 

If you are hearing this statement for the first time, your reaction may be that this state of affairs is 
quite unacceptable. If so... weigh this against the risks implied by a decision not to have nuclear power; 
these latter include cold, starvation and social chaos. 



417 




Figure 1. A CEGB estimate of electricity demand 
until the year 200O-- and how they expect to meet 
it. The energy units have been altered. 



Figure 2. A CEGB estimate of electricity demand 
from nuclear power stations until the year 2000 
(derived from Figure 1) . 



Comments like this tend to portray nuclear systems as the magical source of virtually infinite energy, or at least 
as an incredible untapped source that we cannot afford not to exploit. This report examines such assumptions and 
finds them unproven and probably untrue. Before Britain embarks on expensive nuclear construction programmes, it 
must be clearly demonstrated that these programmes are part of the solution rather than part of the problem. 



(Throughout most of this report, power — in the physicist's sense — is referred to as energy/year. The author 
realizes that he could use the ambiguous word "power", but feels that his meaning is clearer if he does not.) 



2. ENERGY EVALUATIONS OF ENERGY CONVERSION PROCESSES 



Energy conversion processes change a natural energy flow (such as sunlight) or the energy locked in a raw 
material (such as coal) into a form that can be conveniently used by a consumer. These processes require plant tc 
be built, fuel prepared, and output energy distributed to the users. Each of these operations requires inputs of 
energy. The present analysis is concerned with three questions: 

■* How much time does it take before one energy conversion plant has produced more cumulative energy 
than it has consumed? 

♦ How much time does it take before a programme of several such plants has produced more cumulative 
energy than it has consumed? 

♦ What fraction of the energy output of such a programme must be offset against energy needed for 
continual investment in that programme? 

The analysis can be used for any conversion process. In §3 and §4 of this report the analysis will be performed 
for nuclear power stations; but it can be readily applied to other systems such as those that convert North Sea 
oil in situ, or solar or wind energy, to various kinds of useful energy. 



2,1. ENERGY PROFITABILITY OF A SINGLE ENERGY CONVERSION PLANT 



Energy inputs to a plant come in two main categories: investment inputs and process inputs. (A few examples 
of energy inputs to nuclear power that do not, strictly speaking, fit into either category are noted in S4.2.) 
The concepts of investment and running costs are familiar ones in economics. Any business requires an investment 



418 



of capital before it can begin operation, and it also requires expenditure on raw materials or stock throughout 

its period of operation. Energy investment and process inputs are exactly analogous to these economic costs. The 

energy investment includes all energy expenditures, for example, to build energy conversion plants before they can 

start to operate. The process energy inputs are required for the plant to continue in operation, and must be 

supplied during the period of operation. More specifically: 

+ Investment 

Before the plant can give any energy output, the site must be prepared, buildings constructed, machinery 
made, and the initial fuel charge (where applicable) made and supplied. Each of these operations consumes 
energy. Accordingly, during the construction period the plant is a consumer of energy rather than a 
supplier. 

♦ Process inputs 

After the plant has begun production, it still needs energy inputs for maintenance and for preparing further 
fuel. These inputs are considered as deductions from the output of the plant. 

Such a plant is profitable in energy terms when its cumulative output of energy (gross output less process 

inputs) exceeds its total energy investment. A typical plant may take, say, five years to construct; we shall 

assume that the investment is (5 years x P. units of energy/year), P. being the average energy consumed per year 

during the construction period. After construction, a delay of one year is assumed for checking, adjustment, etc 

before the plant is turned on. During its operating lifetime (typically 25 years) the plant will produce an 

average output of energy at a rate P Q /year, Input and output are illustrated in Figure 3 for the construction, 

delay, and operating periods: we assume for convenience that P. and P are uniformly distributed over the 

construction period and the lifetime respectively. 

output 

energy 
P«r 
year 



< years > 

Figure 3. A simple illustration of output energy/year during construction and operation of 
a single energy conversion plant. 




Though the output energy /year is positive as soon as the plant is operating (after six years) , the energy debt 
incurred in its construction will not be offset until after it has been operating for some time. The energy debt 
incurred is 5 x p., and after t years of operation the cumulative output of energy from the plant will be P Q x t. 
Clearly the time at which the plant becomes a net producer of energy is t = 5P./P • If, for example, the output 
energy/year of the plant, P , is twice the energy investment/year, P., then Pq/P^ = 2 and t = 2h years: that is, 
if the plant produces energy twice as fast during operation as it consumes energy during construction, then 
the energy investment will be paid back twice as quickly as it was paid out, or in half the time (five years) 
during which it was paid out. Completing our example, since the period of operation required for the plant to 
recover the energy invested in it is 2"j years, the plant will begin making an energy profit 8>s years after the 
start of its construction. 

Clearly the time that elapses before an energy conversion plant shows an energy profit depends stxongly on 
its ratio P /P.. In §3 this ratio is estimated for six types of nuclear power stations. 



2.2. ENERGY PROFITABILITY OF A PROGRAMME OF PLANTS: A SIMPLE EXAMPLE 

Often more than one plant is required to satisfy the projected demand for energy in the form that the plant 
is to produce. Normally the energy supplier plans a building programme with anticipated demand in mind. (He may 
later have to stimulate enough demand to use the output of his programme — a problem ignored in this paper.) In 
§2.2 and §2.3 we shall explore, for such programmes, the time taken to achieve energy profitability. 

By way of illustration, consider a programme of plants each requiring (as in §2.1) energy investment/year P L 
for five years, then neither requiring nor producing energy for a year, and then producing an output/year P q 
(from which process inputs have been subtracted). Let us also assume that the construction of one plant is begun 
each year. Thus 

in year 1 the number of plants under construction is 1 



419 



- 4 



and so on. Note that after year 5, as one plant is completed, another is begun, so that the number being built is 
constant and equals five. Hence after year 5 the annual energy input is 5 x p^. 

In year 7 the first plant begins operation so that 

in year 7 the number of plants operating is 1 
" " 8 " - " "2 

■ ■ 9 " " " "3 

and so on. Thus the energy output is 1 x p in year 7, 2 x p in year 8, and so on. (After plants reach the end 
of their lifetime — which could be, say, 25 years — one plant will be retired one year as a new one is commissioned, 
so that after year 32 the number of plants operating will be constant and equal 25.) 

We can now build up a table of energy invested and produced per year as the programme progresses (Table 1) : 



(1) 


(2) 


(3) 


(4) 


(5) 


(6) 


(7) 


year 


energy invested 


output energy 


total energy 


total energy 


ratio Pj/Pr 
required if 


ratio PyP; 
required if 




in this year 


in this year 


invested up 


output up to 




(in units of 


(in units of 


to and in 


and in this 


instantaneous 


cumulative 




P i> 


V 


this year 


year (in 


input is to 


energy output 




(in units 


units of P Q ) 


equal instan- 


is to exceed 








of P^ 


taneous output 


cumulative 










in this year: 


energy input 














after this year 


1 


1 





1 









2 


2 





3 





- 


- 


3 


3 





6 





- 


- 


4 


4 





10 





- 


- 


5 


5 





15 





- 


- 


6 


5 





20 





- 


- 


7 


5 


1 


25 


1 


5/1 = 5.0 


25/1 = 25.0 


8 


5 


2 


30 


3 


5/2 =2.5 


30/3 = 10.0 


9 


5 


3 


35 


6 


5/3 = 1.67 


35/6 = 5.83 


10 


5 


4 


40 


10 


5/4 = 1.25 


40/10 = 4.00 


11 


5 


5 


45 


15 


5/5 = 1.00 


45/15 = 3,00 


12 


5 


6 


50 


21 


5/6 = 0.83 


SO/21 =2.38 


13 


5 


7 


55 


28 


5/7 = 0.71 


55/28 = 1.97 


14 


5 


8 


60 


36 


5/8 = 0.62 


60/36 = 1.67 


15 


5 


9 


65 


45 


5/9 = 0.56 


65/45 = 1.44 



TABLE 1. ENERGY INVESTMENT, OUTPUT, ETC FOR A PROGRAMME IN WHICH CONSTRUCTION OF ONE 
CONVERSION PLANT IS BEGUN EACH YEAR. 

(The entry in column 6 for a particular year is the ratio P /P. required for the entry in column 3 
to equal the entry in column 2. Similarly, the entry in column 7 is the ratio P /P. required for 
the entry in column 5 to equal the entry in column 4.) 

Figure 4 shows the variation in energy /year output from such a programme during its early years, assuming 
P /P. = 2. For a single plant, the energy/year output is positive as soon as the plant begins operation, i.e. at 
the beginning of the seventh year after its construction begins; but the effect of putting the plant into the 
programme is to delay for a further year the achievement of positive annual energy output from the programme. 
Table 1 shows that the cumulative energy profit is achieved after 13 years, compared to 8S for the single plant. 




Figure 4. The variation in energy /year (in Table 1, coin 
in which one plant is built each year and P = 2 x p 



3 minus column 2) resulting from a progr 



420 



It is clear from Table 1, and shown further in Figure 5, that the ratio P /P. is very important in determining 
the time that will elapse 

■* before the programme will produce more energy /year output than is required for investment (column 6) and 
+ before the accumulated energy debt is repaid (column 7) . 




30 P /P, 



Figure 5. Plots of the dependence of the time that elapses before (a) the programme produces more energy /year 
output than is required for investment and (b) the accumulated energy debt of the programme is repaid, 
for various values of the ratio P /P., and assuming one plant is built each year, 

A constant rate of construction has been assumed so far in order to provide a simple illustration. Now to 
the real worldj 

2.3. ENERGY PROFITABILITY AND EXPONENTIAL GROWTH 



As a general rule, energy supply utilities aim to supply energy to match expected demand for their product. 
Often these estimates of future demand display a good approximation to exponential growth: the projected demand 
increases each year by a fixed percentage of the demand of the previous year, a process familiar to people who earn 
or pay compound interest on money. Another way to look at exponential growth in demand is to note that if the time 
taken for the demand to double is T , then in each subsequent interval T the demand will double again. In this 
discussion the rate of growth will be characterized by this doubling time T . A good rule of thumb is that T 
equals 69.315 divided by the annual rate of growth in percent. 

Figure 6 compares the curve in Figure 2 (a CEGB estimate of English and Welsh demand for electricity from 
nuclear power stations) with an exponential curve that has T 
until about 1995. 



4rr years; the two curves are nearly identical 




Figure 6. The curve (a) of Figure 2, compared to an exponential curve (b) with T = *r- years. 



421 



(Figure 6 is included only as an example at this stage. In this section we are considering the profitability of 
energy conversion plant and programmes in general; we shall discuss nuclear power specifically in 5 3 and §4.) 

For the moment it is assumed that energy supply utilities treat the energy investment for their plant under 
construction as a demand on supply like any other, even though it is necessary for their own programme to be 
carried out. In other words, the steel, concrete, coal, copper, and other inputs needed to establish a finished 
power station or other energy conversion facility will be treated as an ordinary part of final demand by society. 

The Appendix to this paper analyzes numerically a theoretical programme of plant construction and operation 
intended to produce an exponentially increasing amount of energy output. The analysis shows that the fraction 
of annual energy output whose equivalent is required for investment in the construction of further plant is 
independent of time, if we assume that the operational lifetime of plant is infinite. Figure 7 shows the percen- 
tage of energy output/year that is required (or whose equivalent from another previously uncommitted source is 
required) for investment, for various values of P /P. characteristic of the individual plant. The plot has been 
repeated for doubling times T = 2, 3, 4, 5, and 8 years. 



(For this graph and the others in this section, save the noted exceptions, it has been assumed that the plant 
takes five years to build, that there is a one-year delay before it is commissioned, that its average energy 
investment/year during construction is P , and that its average energy output/year throughout its operational 
lifetime is P .) 



investment 
as a °/o of 




Figure 7. The percentage of energy output that would be required for investment, plotted against the ratio 
P /P. for plant used in exponential programmes with characteristic doubling times T = 2, 3, 4, 5, and 8 years. 
(As will be shown later, it is difficult in practice to achieve P /P. greater than about 2 or 3 with nuclear 
power stations.) Values greater than 100% on the vertical axis correspond to values of P /P. and T for which 
deployment of plant requires a continuing external subsidy. The curves are all calculated from Equation A. 6 
in the Appendix. 



The effect of considering the lifetime of each plant is to increase the output-to-investment fraction shown 
in Figure 7 during the period of replacement. The extent of this increase is calculated in the Appendix and illus- 
trated in Figure 8. Here the percentage of output energy required for investment is plotted against Pq/P,, both 
with (dashed lines) and without (solid lines) retirement of plant at the end of its lifetime. For a given ratio 
P /P., the percentage of output energy that is equivalent to the required investment can be read from the graph for 
programmes with doubling times T = 2, 3, 4, 5, 6, and 10 years, where 

+ the lines marked (a) give the exact percentage for the operational lifetime until T = 25 years, which is 
assumed to be the life of individual plant. The percentage is also approximately correct for T > 25 if 
T D = 2 or 3. 
* the lines marked (b) give the exact percentage for the period from T ■ 25 to T » SO years. The percentage 
is also approximately correct for T > 50 if T = 4, 5, or 6. 



422 



Investment 
as . % 




'J'x 



Figure 8„ The percentage of energy output equivalent to the energy that would be required for investment, 
plotted against the ratio P /P. (this time using a reciprocal scale) for plant used in exponential programme 
with characteristic doubling times T = 2, 3, 4, 5, 6, and 10 years. The dashed lines show how the percen- 
tage will change if the programmes retain these rates of growth until the first (b) and second (c) replace- 
ments of obsolete plant occur (see text). The dashed lines are omitted in the cases T = 2, 3 because they 
would fall essentially on top of the solid lines, which show the effect of assuming infinite plant lifetime. 
All the dashed lines shown are valid for times outside of short periods when early transients in the pro- 
gramme are "echoed" (Appendix, Equation A. 6) , but the correction in these periods would not greatly change 
the results shown. 



■* the line marked (c) gives the exact percentage for the period from T = 50 to T = 75 years. The percentage 
is also approximately correct for T > 75 years. 

These percentages are applicable during the period of operation of the programme. However, before the pro- 
gramme has produced any outputs, energy has been consumed to construct initial plant. One question of interest is 
how much time is taken before the programme produces more cumulative energy than has been used in this initial 
construction period. Figure 9 plots against P /P. the time required for various programmes (with doubling times 
T = 2, 3, 4, 5, 6, and 10 years) to make an energy profit, assuming that plant life is infinite. (If plant life 
were assumed finite, the programmes shown would gradually become less energy-profitable, in term of percentage of 
output to investment (Figure 8) , as successive generations of retired plant had to be replaced. The position would 
of course be more complex if new types of plant were brought into use so that P /P. varied in time; the present 
analysis is merely illustrative and is not designed to deal with such complexities.) 

Consider three examples. First, suppose that an energy supply utility has a conversion plant whose lifetime 

is 25 years and whose annual energy output-to-investment ratio P /P. is three. If the utility is planning to use 

such plant in a programme designed to produce an output that doubles every five years (T = 5), Figures 7, 8, and 

9 show that 

-»■ from the time the first plant begins operating until 25 years from the beginning of its construction, the 
energy required for ongoing construction of plant will be equivalent to 38.5% of the programme's output/year. 

* from then on the percentage will be approximately 40%. 

♦ the time taken before the programme has produced more cumulative energy than it has consumed will be about 
8S years. 

If the plant has a ratio P /P. =2, then 

■+ for the productive period of the first 25 years (i.e. 6 < T < 25 years) the percentage of output that must 
be offset against investment will be 57.5%. 



■+• from then on the percentage will be 60%. 

+ the time that will be taken before the progr 
If the plant has a ratio P /P. =1, then the progra 
and will never make a profit. 



5 has made a "profit" is about 13*j years, 
will always consume more energy than it produces in output 



423 



yart 

to 
r»poy 

d.bt 




8 Kl', 



Figure 9. The time that elapses before the accumulated energy debt is repaid for plant used in exponential 
programmes with characteristic doubling times T = 2, 3, 4, 5, 6, and 10 years and with various P /P.. 
(Calculated from Equation A o 10 in the Appendix.) 

For these three examples, Figures 10, 11, and 12 respectively show the net and gross output energy /year 
from the beginning of construction of the first plant. Each graph shows, in slightly simplified form, discontinui- 
ties in output corresponding to (a) the operation of the first plant and (b) the beginning of construction of addi- 
tional plant to replace that which becomes obsolete. The shaded area shows the energy/year required for construc- 
tion of plant as a function of time, and the black area shows the energy /year required to replace obsolete plant. 

If investment energy were mistakenly assumed to be negligible in these examples, then the energy /year made 
available to society by the programme would be approximately that shown by the upper (gross energy) curve. But 
the energy /year actually available for purposes other than building plant is shown by the curve forming the 
bottom edge of the shaded and black areas. (For exact values, use Equation A. 7 in the Appendix.) 

For the case where the ratio P /P. = 1, there is clearly no energy left — indeed, for the programme to continue, 
it must receive an ongoing energy subsidy from some other source. let a single plant with this P/P-- = 1 is a 
perfectly profitable plant which produces in its operational lifetime five times* more energy than was required 
in its construction. It is the speed and pattern of the programme of construction and operation that can turn a 
respectable energy producer into just another consumer. 

This discussion has assumed that, in devising programmes of plant construction and operation, the supply 
utility has taken into account the energy demand of the programme itselfT (Essentially, in this treatment we 
have been concerned with the rate of increase of demand, not with its magnitude.) If the utility has not properly 
included this investment demand as part of total demand, then the programme must of course be augmented accordingly 
by building the equivalent of more than one plant in place of each one that had been planned. This will increase 
the output of energy available for societal use, but will also increase the demand for energy to be invested in new 
plant. Moreover, there is no way in which a non-profitable programme can be made profitable: the best that can be 
achieved is not to start it at all. 

These examples assume a programme with a characteristic doubling time T = 5 years. Figure 13 plots the 
energy/year required for investment (as a percentage of annual output) against various doubling times for different 
values of the ratio P^/P,- For these plots plant lifetime has been assumed infinite, construction time 5 years. 



* Eight times more if the plant lifetime is not 25 years but 40 as is commonly assumed in the USA (and if the 

construction time remains five years) . 
t I.e. it is assumed that the utility considers final demand by consumers other than the nuclear industry to be 

defined by the lower, not the upper, boundary of the shaded and black areas in Figures 10-12. 



424 



N «Po 


i 
















28 


















24 




P c /P| = 3 














20 


















16 


















12 


















8 


















4 




















4 










. 


years 


, 






8 12 


16 


20 














N 0* P o 


















28 


















24 




P /P, = 2 














20 


















16 


















12 


















8 


















4 




















-4 


4 | 












yton 




uTiihrrrftfAtA 


8 12 


16 


20 


24 


28 


32 






Figures 10, 11, and 12. Net and gross energy/year output from a programme designed to produce a gross output 
increasing exponentially with a doubling time T D = 5 years. The construction time for each plant is assumed 
to be five years, and one further year's delay is assumed before commissioning. The shaded area represents 
the energy required for investment in construction of new plant, and the black area represents the energy 
required for further construction to replace obsolete plant after a lifetime of 25 years. N e is the number of 
plants whose construction is undertaken in the first year; P is output/year from one plant. (As will be seen 
later, the five-year doubling time assumed here is longer thfan that of virtually any current nuclear programme. 



425 



elapsed before the cumulative energy debt is repaid for the 



values of T„ and 



Figure 14 plots the 
P /P. used in Figure 13. 

Figures 13 and 14 relate programme profitability to T and P /P.. For types of plant representative of the 
ratios P /P. considered, the faster rates of growth shown are clearly unrealistic if intended to increase rapidly 
the supply of energy; for though each plant is individually energy-profitable, a programme of such plant can 
become a major net energy consumer if the rate of deployment exceeds certain limits which we shall now consider. 

Figure 15 shows these limits in a way that energy policy-makers might find useful. For varying P /P., Figure 
15 plots doubling times which would result in a given percentage of energy output being offset against investment 
energy in the programme. For example: 

♦ the policy-makers wish to build a programme of plant whose P /P. is two. 

♦ they are prepared to invest in the programme up to the equivalent of 75% of its output energy. 

Then from Figure 15 the fastest sustained exponential rate of construction which would give this result is 
that with a doubling time T = 4.4 years. 

If the percentage of output they were prepared to invest were only 25%, then the minimum sustained doubling 
time allowable would be 8.8 years. 





10 12 yr 



Figure 13. Energy/year required for investment 
(as a percentage of output/year) ve exponential 
doubling time T D , for various values of P Q /Pi 
(and infinite plant lifetime) . (Calculated 
from Equation A. 6 in the Appendix.) 



Figure 14. Time before cumulative 
energy debt is repaid, with assumptions 
as in Figure 13. (Calculated from Equation 
A. 10 in the Appendix.) 



This report assumes that individual plant takes five years to build and a further year to commission. Since 
P. is defined as the average energy investment/year during the construction period, the ratio P /P . is affected 
if the construction time changes. If the construction time has some value T other than five years, then the 
corresponding new value of P Q /P i can be found by multiplying the P Q /P i shown in the graphs by the quantity 5/T c . 

For purposes of demonstrating the effect of a change in the assumed construction time, let us take as an 
example a programme of plants which have a ratio P /P . = 2, a construction time initially taken to be five years, 
and a characteristic doubling time T D = 5 years. What, then, will be the percentage of energy output required 
for investment for plants under construction, and how much time will elapse before the accumulated energy debt 
is repaid, if the construction time is something other than five years? 



68-391 O - 76 - 28 



426 




Figure 15. Exponential doubling times T which, for given values of P /P., will require that the 
indicated percentages of programme output/year be offset against investment in the programme. (E,g. 
the curve labelled "25%" would be appropriate if it were desired that 75% of the programme output 
be available for purposes other than investment in new plant.) Plant lifetime is assumed to be infinite. 
The shaded area will be explained in 56. 



This percentage (Figure 16) and time (Figure 17) are plotted against a range of construction periods. The 
effect of altering the delay time between the end of construction and the commissioning of individual plant is also 
shown, in each of the plots, by dashed curves for a delay time of two years (a) and zero years (b) along with the 
assumed delay time of one year (the solid curve). Figure 16 (but not Figure 17, for which the result would be far 
off-scale) also shows how the plotted sensitivity increases steeply with doubling times T shorter than five years. 

(Note that the shorter the doubling time T D , the 
greater the effect of delays in construction and 
commissioning. ) 



£ 20 




construction time 



construction 



Figure 16. Percentage of output/year required 
for investment vs construction time, for 

delay time of 1 year , 2 years - (a), 

and O years - - - -(b). (Equation A. 6, Appendix.) 



Figure 17. Time elapsed before the cumulative energy 
debt is repaid vs construction time, for delay time 

of 1 year , 2 years - - - -(a), and years 

- - - -(b). (Equation A. 10, Appendix.) 



These graphs show that the shorter the construction period and time delay are, the more profitable in energy 
terms is the programme and the shorter the time required to repay its energy debt. Thus if, with T D = 5 years, the 
above analysis had correctly estimated the total energy required for investment in a plant as E^ (= 5 x p^) but 
had wrongly assumed that the construction time is five years when it was really four, then investment energy as a 



427 



percentage of output energy would be 53% instead of 57%. The change in the time that would elapse before the 
energy debt was repaid would be from 12.2 years (construction time 5 years) to 10.5 years (construction time 4 
years). If, on the other hand, the construction time had been underestimated, the effect would be slightly 
larger in the other direction. For T < 5 years, the sensitivity would be greater than Just stated (Figure 16). 
This section has discussed the implications of exponential growth programmes on the basis of the analytical 
treatment presented in the Appendix. This in turn has assumed that the number of operating plants is given by a 
smooth exponential function. In fact this condition can only be approximated in reality; but the author has 
checked the analytic approximation by unsmoothed simulation and has found it to be quite close, particularly 
after the first few years. 

2 A. SUMMARY 

For the two types of programmes considered — linear (§2.2) and exponential (§2.3) — the above analysis shows 
that energy profitability depends strongly on the ratio P Q /P i , where P. is the average annual energy investment 
in the construction of a single plant and P is the average annual output of energy from that plant throughout 
its operational lifetime (less its process inputs) . 

Quite respectable types of plant which produce far more energy in their lifetime than is required in their 
construction can collectively become net energy consumers as part of a programme. At a time when energy production 
is requiring increasing inputs of energy, care must be taken if programmes are not to become significant consumers 
of the energy they produce, or even of further subsidies besides. 

The next section will estimate the ratio P /P. for six types of nuclear reactor; §4 will discuss the implica- 
tions of using such reactors in the sorts of programmes we have considered here. 

3. STATIC ENERGY ANALYSIS OF NUCLEAR REACTORS 

The analysis just described (§2) is rather simple; the difficulties begin as one tries to quantify the actual 
energy inputs and outputs for a singular and particular energy conversion system in isolation. This is called a 
static analysis to distinguish it from the dynamic analysis that considers inputs and outputs for a programme as a 
function of time. 

This section summarizes a static analysis performed mainly by Chapman and Mortimer for six types of nuclear 
power stations. For convenience, each power station is considered to be designed to produce 1000 MW electrical 
output (net of electricity consumed by pumps, controls, etc at the power station itself). 

3,1. ENERGY INVESTMENT IN CAPITAL PLANT 

The amount of energy required to build a nuclear power station is not now known precisely. Until detailed 

engineering analyses are available in a year or two, this energy investment must be estimated from aggregated 

financial data. To do this it is necessary to separate the capital costs of the power station into four parts: 

+ electrical machinery, including the generator set, power transformers, control and switchgear, and 
distribution links to the grid; 

♦ buildings and services, including the site itself, office blocks, buildings to house equipment, 
cooling towers, service roads, and provision of an adequate water supply; 

* the initial core assembly (§3.2), which is an initial supply of fuel amounting to about 1/8* of the 
total fuel which the reactor will consume in its lifetime (for a nuclear power station, unlike a 
coal-fired power station, this initial fuel must be supplied before operation can begin, and is 
therefore an investment input rather than a running or process input) ; and 

-»■ the nuclear reactor and steam system, including containment devices, safety and control systems, 
steam circuits, and heat exchangers. 

Electrical machinery and the buildings and services are significant cost items for both conventional and nuclear 

power stations. The initial core assembly and the nuclear steam supply system (notably the reactor) are, however, 

special requirements of nuclear power systems, and reactors tend to cost more than equivalent fossil-fuelled 

boilers. 

4 
The CEGB have published an approximate breakdown of capital costs for five types of nuclear power stations, 

showing separately the costs for total construction and for the nuclear steam system. By comparison with the 

known capital costs for electrical equipment for coal-fired power stations, Dr Chapman has derived the division 



Assuming a 25-year lifetime; the "-1/8 would be 11/13 if the lifetime were assumed to be 40 years according 
to US practice. 



428 



of the remaining costs of nuclear power stations into costs for electrical equipment and for buildings and servi- 
ces. The breakdown into the various categories is tabulated below for the five types of reactor described* by the 

4 
CEGB and for the HTR as described by TNPG*. 



TABLE 2. CAPITAL COSTS (E/kW installed electri 


;al output capacity; 1973 price 


levels) 


reactor type 


electrical 


buildings 


nuclear steam 


initial 


TOTAL 




equipment 


& services 


system 


fuel 




Magnox 


52 ± 10 


73 ± 5 


116 ± 2 


15 ± 2 


256 ± 2 


SGHWR 


52 t 10 


31 + 5 


67 ± 2 


18 ± 2 


168 ± 2 


PWR 


52 ± 10 


30 + 5 


50 + 2 


14 + 2 


146 + 2 


AGP. 


52 ± 10 


30 ± 5 


89 ± 2 


19 ± 2 


190 + 2 


CANDU 


52 i 10 


31 ± 5 


67 ± 2 


12 ± 2 


162 ± 2 


HTR 


52 ± 10 


30 ± 5 


60 ± 2 


18 + 2 


160 + 2 



(Magnox = C0 2 -cooled, graphite-moderated reactor fuelled by Magnox-clad natural uranium 
metal; SGHWR = steam-generating heavy-water-moderated reactor; PWR = pressurized-water- 
cooled-and-moderated reactor; AGR = advanced gas-cooled reactor, C02-cooled, graphite- 
moderated, U02-fuelled; CANDU = Canadian deuterium-uranium reactor, heavy-water-modera- 
ted, heavy-water-cooled, natural-uranium-fuelled; HTR = high-temperature reactor, graph- 
ite-moderate, helium-cooled, ceramic-particle-fuelled) 



From statistical surveys of the total direct and indirect use of fuels, the average ratio of energy inputs/E 
value for all UK industries has been computed * for 1968= The energy consumed in the preparation of the initial 
fuel charge (§3.2) and of refuelling loads (§3.3) varies considerably with the ore grade assumed, and also depends 
on whether the reactor uses enriched or natural uranium, so it must be considered separately, as must the energy 
requirement for heavy water (§§3.2-3). For the other categories, the average UK ratios are used; corrected 
to 1973 £, these are: 

•+ for electrical equipment, 33 kWh(t)/£; 

■* for buildings and services, 28 kWh(t)/£; and 

■* for the nuclear steam system, 30 kWh(t)/£. 
The significance of errors associated with the use of these national industrial averages will be discussed later. 



3.2, SPECIAL INVESTMENT INPUTS AND SOME CORRESPONDING PROCESS INPUTS 



3.2.1. URANIUM FUEL 

The initial charge of uranium fuel represents a major energy investment in the construction of a nuclear reac- 
tor. The energy needed to prepare reactor fuel depends on the grade of ore mined and on whether the reactor uses 
enriched or natural uranium. For this reason two limiting grades of uranium ore have been considered here. The 
high-grade ore (0.3% U,0_ by weight) is the average grade of ore being currently mined in some US mines. The 
second ore considered is Chattanooga Shale with an average grade of 0.007% U,0_ by weight: this source is chosen 
not because it will necessarily be typical of future low-grade uranium sources, but because it is the low-grade 
source for which the most detailed mining-engineering assessments are available. 

For each of these grades of ore, the energy needed to process the ore for use in the reactor has been calcu- 
lated by considering separately the following processes: 

♦ mining and milling to extract U,0„; 
+ converting the U,0. to UF_; 

♦ enriching the uranium in the UF-; and 

♦ converting the UF to the solid fuel material and fabricating it into fuel elements. 

The detailed calculations are not included in this report, but a version sufficiently detailed for most purposes 
is available and is based on published data 

The energy needed to extract a tonne of uranium in a useful chemical form from uranium ore is shown in Table 3. 



TABLE 3. 

process description 



ore from conventional sources I ore from Chattanooga Shale 





(0.3wt% U 3 8 ) 


(0.00?Wt% UjOg) 


mining & milling 
conversion to UFg 


0.265(t) 
0.016(e) + 0.054 (t) 


18.63(t) 
0.016(e) + 0.054(t) 


TOTALS 


0.016(e) + 0.319(t) 


0.016(e) + 18.68(t) 



These data were slightly modified after communication with The Nuclear Power Group (TNPG) , who also supplied 
the HTR data shown. Many US sources would give considerably higher estimates of capital costs. 



429 



Enrichment 

The SGHWR, PWR, AGR, and HTR reactors use enriched fuel. Uranium naturally occurs as a mixture of two main 
isotopes, 238 u (99.29%) and 235 u (0.71%). The enrichment process partially separates these isotopes so as to raise 
the proportion of 235 U in the enriched uranium produced for use in fuel. The amount of "separative work" required 
for this separation depends on the degree of enrichment (.i.e. the percentage of 235 u required in the fuel). The 



U (the "tails"), though partially depleted in 235 U, unavoidably contains some small residual percen- 
1.14 



separated 

tage of 23:> U (the "tails assay"), typically 0.25%*"". This paper will assume the 0.25% value throughout. 

Table 4 shows the enrichment energy required per tonne of enriched uranium in fuel and the natural uranium 
input to enrichment required per tonne of enriched uranium in fuel for the SGHWR, PWR, AGR, and HTR. 



fuel for 


enrichment 


tonnes natural uranium 
input per tonne enriched 
uranium output^- 


energy required (10 kWh) per 
tonne enriched uranium output 1 '^ 


17,18 


AGR initial core 
" refuelling 


2.45% 
2.60% 


4.78 
5.11 


6.53(e) + 0.259(t) 
7.26(e) + 0.288(t) 




SGHWR (both) 


2.10% 


4.24 


4.96(e) + 0.197(t) 




PWR initial core 
" refuelling 


2.70% 
3.11% 


5.33 
6.22 


7.74(e) + 0.307(t) 
9.78(e) + 0.388(t) 




HTR initial core 
" refuelling 


6.50% 
10.00% 


13.59 
21.20 


27.30(e) + 1.083(t) 
46.17(e) + 1.832(t) 





Energy inputs for different reactors depend on the net energy output, initial fuel inventory, annual 
fuel charge, levels of enrichment, etc. No two authorities agree on typical values for these parameters 
for any given type of reactor. This analysis has used those data from the Directory of Power Reactors 
which are most nearly consistent with those in the Nuclear Power Index Jff . Both sources give details of 
actual reactors which may not be fully characteristic. Wherever possible, model reactors for this analysis 
resemble those given by OECD^ 4 or by UKAEA and TNPG in private communications to Dr Chapman. The reactors 
chosen are Oldbury A (Magnox) , Hunterston B (AGR) , and Pickering (CANDU) . SGHWR design characteristics are 
as published by Moore et al , and HTR characteristics are from TNPG (private communication to Dr Chapman). 
The data for the PWR vary greatly for different reactors, and Chapman^ cites four examples; the data used 
here are for the most energy-profitable of these (Shearon Harris) . Different choices of data are justifi- 
able but lead to substantially identical conclusions. 



Fabrication 

Once the UF,- has been enriched (as and if required) it must be converted to UO2 or the like and then 



11 



fabricated into fuel elements. The energy required (10 kWh/tonne enriched uranium) for these processes is 
0.048(e) + 0.032 (t): energy required in the form of electricity (e) has been left separate from other energy 
inputs which can be converted to thermal equivalents to give the (t) or kWh(t) total. The electricity required 
to process the uranium needed for refuelling can be directly subtracted from the reactor output: in this case 
kWh(t) and kWh(e) are assumed equivalent. For the initial fuel charge, however, the electricity itself has an 
energy requirement arising from the present efficiency of conversion of primary energy (mainly from oil or coal) 
to electricity. For the initial fuel charge only, this efficiency is taken into account by using a 25% conversion 
factor, so that 1 kWh(e) becomes 4 kWh(t). 

Using this conversion factor and the energy requirements noted above, the energy needed to provide the 
initial fuel charge for a 1OO0 MW(e) reactor is: 



reactor 


fuel in core 
(tonnes U) 


% enrich- 
ment 


requirement 
of natural 


ENERGY REQUIREMENT (10 


5 kUh) 




type 


mining, milling, 


enrich- 


fabri- 


TOTAL 








U (tonnes)* 


and conversion 


ment 


cation 














0. 3%ore 


. 007%ore 






0. 3%ore 


0.007iore 


Magnox 


973 


natural 


973 


373 


18244 





218 


591 


18462 


SGHWR 


160 


2.1 


657 


252 


12319 


3206 


36 


3494 


15561 


PWR 


87 


3.2 


463 


177 


8681 


2719 


19 


2915 


11419 


AGR 


195 


1.9 


932 


357 


17475 


5154 


44 


5555 


22673 


CANDU 


182 


natural 


182 


70 


3412 





41 


111 


3453 


HTR 


22.7 


6.5 


315 


120 


5906 


2504 


5 


2629 


8415 



a The requirement of natural uranium is the amount that must be mined, milled, etc to be processed into the 
required amount of fuel; e.g. 657 tonnes of natural uranium is needed to produce 160 tonnes of uranium 
enriched to 2.1% with 0.25% tails assay. 

N.B.: see the note to Table 4 regarding data. 



3.2.2. HEAVY WATER 

The Heavy Water Division of Atomic Energy of Canada Ltd has estimated that the energy required to make a 

tonne of heavy water is 0.65 * 10 kWh(e) ♦ 6 « 10 kWh(t). The CANDU reactor uses heavy water as a moderator 

20 
(0.3 tonne/MW(e)) and as a coolant (0.4 tonne/MW(e)) ; the SGHWR reactor uses heavy water as a moderator only 

(0.25 tonne/MW(e) )". The energy requirements for the heavy-water inventories of reactors of 1000 MW(e) installed 



430 



electrical output capacity are thus 6020 * 10 kWh(t) (CANDU) and 2150 * 10 kwh(t) (SGHWR) . As noted below 
(§3.3), small process inputs of heavy water are also required to make up losses during operation. 

3.2,3. OTHER INPUTS 

The other inputs are estimated using the capital-cost breakdown and kwh(t)/£ multipliers discussed earlier, 
and are incorporated into Table 5 (§3.3). 

Because of the specialized nature of the technologies and the special materials used in building nuclear 
reactors, aggregated estimates of energy investment in reactors are liable to substantial error; but for the 
next year or two no better data will be available. Ideally the special investment inputs to nuclear reactors 
should be considered separately in full engineering detail. Dr Chapman intends to do this for fuel cladding 
(e.g. hafnium-free zirconium), moderators and control rods, heavy water, pressure vessels, and perhaps other 
inputs (§4.2). 

3.3, OUTPUT OF ENERGY FROM A NUCLEAR POWER STATION 



A nuclear power station designed to have 1CO0 MW installed electrical output capacity does not in fact supply 
1000 MW-year of electricity per year to final consumers. The average energy /year supplied is much less and depends 
on the capacity factor, the distribution losses, and the use of electricity by the CEGB itself. 

The working life of a power station is widely accepted in the UK as being 25 years, and the present analysis 

has assumed that 

7 4 

■* the average capacity factor of a nuclear power station over its lifetime is 62%, nearly the 64% assumed 

by the CEGB (capacity factor is the amount of electricity sent out per year, expressed as a percentage 
of the amount which would be sent out per year if the station operated continuously at its full design 
rating) ; and 

+ total losses of electricity between the power station and the consumer amount to 11.25% of the amount sent 
out. This is derived as follows. The distribution losses and use of electricity by the CEGB'^ accounts for 
15% of all electricity generated. Of this 7.5% is classed as distribution losses. The remaining 7.5% is 
used "by the CEGB and area boards in offices, showrooms, workshops etc". Half of this latter 7.5% will be 
assumed to be a loss incurred in activities that support electricity generation; the other half, used in 
CEGB offices, area board showrooms, etc, will be ignored here as an irrelevant overhead. Adding 7.5% 
distribution losses to 3.75% CEGB use yields a total loss of 11,25%. 

Taking these assumptions into account reduces the effective usable power output of a 1CO0 MW(e) nuclear power 

station to 550.25 MW ( = 550.25 x 10 kw) . The annual energy output of such a plant is thus 4820 « 10 kwh. 

(Throughout this paper it is assumed that a year contains 8760 hours, rather than the 8766 that would be needed 

to take account of the extra day in each leap year.) 

To be offset against this annual energy output is the energy required* to process 

+ the heavy water required to make good a loss of 0.7%/year for CANDU and a loss of 0.5% for SGHWR j the 

42 x lo 6 kWh/year and 8 



energy needed for this replacement 



■*■ the uranium fuel required to refuel the reactor. The ann 
needed to produce it are tabulated below. (See the note 



10 b kWh/year respectively. 

L tonnage of fuel required and the energy 
Table 4 regarding data.) 



reactor 


reload fuel 
(tonnes U/ 


% enrich- 
ment 


requirement 
of natural 


ANNUAL ENERGY 


REQUIREMENT (10 6 kWh/year) 


type 


mining, milling , 


enrich- 


fabri- 


TOTAL 




year} 




U (tonnes/ 


and conversion 


ment 


cation 








year) 


0. 3%ore 


. 007%ore 






0. 3%ore 


0. 007%ore 


Magnox 


187.0 


natural 


187 


62 


3500 





16 


78 


3516 


SGHWR 


31.4 


2.10 


133 


45 


2487 


162 


2 


209 


2651 


PWR 


17.4 


3.11 


108 


36 


2020 


178 


2 


216 


2200 


AGR 


30.0 


2.60 


153 


51 


2861 


226 


2 


279 


3089 


CANDU 


67.0 


natural 


67 


22 


1257 





6 


28 


1263 


HTR 


6.8 


10.00 


147 


49 


2755 


325 


0.5 


375 


3081 



The total energy requirement/year of processing the replacement fuel has to be subtracted from the gross 
output of a nuclear power station to give its output energy/year, P . Table 5 shows this net output; the average 
energy investment/year throughout the construction period, P.; and the resulting "power ratio", P / P j» Table 5 
also shows the "energy ratio", E /E., where E is the total lifetime output of each plant (net of process inputs) 
and E. is the total energy investment in each plant. Clearly E /E . = (P /P.) x (plant lifetime / construction 
time); it is assumed here that the lifetime is 25 years and the construction time 5 years, leading to E Q /E i 
equalling 5 x p /p . „ 



For each of these energy requirements, the energy needed per tonne of product is as used in Table 3, §3.2,1, 
but the electrical energy required is offset directly against the output of the power station without the 
use of any conversion factor. 



431 



TABLE 5? AVERAGE ENERGY INVESTMENT/ YEAR P., AVERAGE ANNUAL ENERGY OUTPUT (LESS PROCESS INPUTS) P 
POWER RATIO P /P. , AND ENERGY RATIO E /E. , BOTH FOR FUEL FROM 0.3% URANIUM ORE AND FOR FUEL FROM 
0.007% CHATTANOOGA SHALE (numbers in italics). 



reactor type 


P- h 

CO 6 kwhh)/yr) 


6 P o 
(10 b kWh(t)/yr) 


power ratio F ' JP . 
& error estimate d 


energy ratio F/E . c 
& error estimate^ 


Magnox 


1566 509? 


4734 


1204 


3.02±0.6 0.26+.04 


15.1±3 1.28±0.2 


SGHWR 


2048 4461 


4603 


2261 


2.25±0.4 0.511.08 


11.3±2 2.55±0.4 


PWR 


1394 3095 


4604 


2620 


3.30+0.6 0.8S±.16 


16.5±3 4.23±0.8 


AGR 


2154 5578 


4541 


1731 


2.11±0.4 0.31+.06 


10.5±2 1.56±0.3 


CANDU 


2145 2813 


4750 


3151 


2.21+0.4 1.25±.24 


11.1+2 6.24±1.2 


HTR 


1397 2554 


4445 


1739 


3.18+0.6 0.68±.12 


15.8±3 3.30±0. 6 



See the note to Table 4 regarding data. 

The average energy investment/year throughout the construction period, P^, is calculated on the 
assumption that the construction period for each reactor is five years. 

Chapman and Mortimer discussed their results in terms of this energy ratio, and assumed, as has 
this paper, that the lifetime of each reactor is 25 years. 

In making this analysis, assumptions have had to be made and conventions adopted. Because precise energy- 
requirement data are not now available, some estimates have had to be based on aggregated financial data. 
Dr Chapman estimates the errors inherent in such methods to be ±20% for energy requirements estimated from 
financial data, ±10% for the analysis of energy requirements of uranium processing, and ±5% for the net 
energy estimation (before subtracting the energy required to prepare replacement fuel charges) . Owing to 
the varying importance of the different types of energy requirements for each system, these sources of 
error will not make the same relative contributions to total error. Dr Chapman's estimates of total 
error for the ratios Pq/Pj^, and the corresponding (5x larger) estimates of total error for the ratios 
E /E i , are shown. For e.g. the Magnox reactor using fuel prepared from 0.3% ore, the ratio P /Pi is 
calculated as being 3.02 with an error estimate of ±0.6. This means that we can only confidently say 
(subject to the caveat below) that the power ratio lies between 2.42 and 3.62. It is hoped that further 
analysis will narrow the error limits. Meanwhile, however, the results are not precise enough to permit 
detailed comparison of reactor types. 

These error estimates assume that all significant energy requirements have been identified and considered, 
and that the data used are a fair representation of the types of reactors shown. The error estimates are 
also not intended to cover possible omissions from the analysis; these are discussed in §4.2 and may 
significantly alter the energy and power ratios shown here. 



Bearing in mind the possible errors but not the possible omissions (14.2) of this analysis, we can draw the 

following conclusions: 

+ For high-grade ore (0.3% U30 8 ) , it is impossible to distinguish significantly between the six types of 
reactors in terms of energy profitability. 

♦ The energy ratios of all the reactors are much reduced if their fuel has to be processed from low-grade 
ore such as Chattanooga Shale. In this case the energy requirements for mining are very significant. 

Using the power ratios determined in this section, we can now examine the implications of incorporating the 

various types of reactors into the types of programmes discussed in §2. 



*k DYNAMIC ENERGY ANALYSIS OF NUCLEAR POWER PROGRAMMES 
4,1. ENERGY PROFITABILITY OF EXPONENTIAL NUCLEAR POWER PROGRAMMES 



Estimates of the ratio P./Pj f°r six types of reactors have been given in §3. This section will consider 
the implications of using such reactors in a programme which aims to satisfy an exponentially increasing demand — 
a problem discussed in general terms in §2. 

If high-grade ores are processed to fuel the reactors considered, the ratio P /P. is found (§3) to range from 
2.11 to 3.30. Table 6 shows, for these values, how much continuing energy investment is required by the programme 
(expressed as a percentage of the programme's output), and how much time elapses before the initial energy debt is 
repaid, for programmes designed to achieve an exponential growth in gross energy supply with characteristic 



doubling times T 



>, or 6 years. Such rates are chosen here because they seem representative of the thinking 
>f the CEGB (see Figure 6), Sir John Hill (Chairman of the UK Atomic Energy Authority), and Bainbridge and 



Beveridge o 
P o /P i 


C the UKAEA. 

T D = 4 years 


Tn = 5 years 


T n = 6 years 


% output to 
investment* 


years to 
pay debt 


% output to 
investment* 


years to 
pay debt 


% output to 1 years to 
investment* pay debt 


2.11 
3.30 


78% 
50% 


14.7 
10.0 


54% 
35% 


11.7 
9.1 


42% 10.7 
27% 1 8.7 



TABLE 6. ENERGY INVESTMENT IN AN EXPONENTIAL PROGRAMME AS A PERCENTAGE OF ENERGY OUTPUT FROM THE SAME 
PROGRAMME, AND TIME REQUIRED BEFORE THE PROGRAMME'S OUTPUT HAS REPAID THE CUMULATIVE INVESTMENT ENERGY, 
FOR THE BEST (V Q /V i - 3.30) AND WORST (Po/Pi = 2.11) REACTORS USING FUEL PROCESSED FROM HIGH-GRADE ORE 
a These percentages apply from the start of operation until one must start to build new reactors to replace 
obsolete ones (after 25 years). The increase arising from such replacement is small (see the Appendix). 



432 



If low-grade ores are processed, then each reactor considered in isolation will produce more energy in its 25- 
yaar lifetime than was required in its construction. Yet considered dynamically in the same exponential programmes 
Just described, the reactors will perform as in Table 7, which again shows the best and worst ratios P /P . 



v o /? i 


T r - 4 


sears 


Tn= 5 


years 


Tn m years 




X output to 
investment* 


years to 
pay debt 


% output to 
investment* 


years to ! 
pay debt j 


% output to I years to 
investment* \ pay debt 


0.26 
1.25 


630 
131 


never 
never 


442 
92 


never 

24 


337 1 never 
70 1 16.5 



TABLE 7. ENERGY INVESTMENT IN AN EXPONENTIAL PROGRAMME AS A PERCENTAGE OF ENERGY OUTPUT FROM THE SAME 
PROGRAMME, AND THE TIME REQUIRED BEFORE THE PROGRAMME'S OUTPUT HAS REPAID THE CUMULATIVE INVESTMENT ENERGY, 
FOR THE BEST (P /Pi = 1.25) AND WORST (P^Pi = 0.26) REACTORS USING FUEL PROCESSED FROM CHATTANOOGA SHALE 
a These percentages apply from the start of operation until one must start to build new reactors to replace 
obsolete ones (after 25 years) . The increase arising from such replacement is small (see the Appendix) . 



In §2, the energy profitability of programmes was considered for energy conversion facilities in general. 

Figures 7 and 8 plotted investment as a percentage of output for varying P /P.. The values of this ratio estimated 

for nuclear reactors in §3 are such that marginal increases in P ± (or decreases in P ) will significantly reduce 

the energy /year available for uses other than investment. 

Increases in the actual value of P. can occur because of declining ore-grade, more stringent safety requlre- 
25 
ments , increased security measures, and the like. Decreases in the actual value of P can occur because of 

reduced station output rating (e.g. because of corrosion or revised safety analyses), reduced capacity factor 

(.e.g. because of unexpected aging problems ), and the like. One or more of these factors could turn a 

profitable programme into an energy consumer. 

Of course, technological advance might lead to increased P /P., but the resulting improvement in programme 
profitability would be much smaller than the loss due to an equivalent decrease, owing to the concavity of the 
curves shown in Figure 7. In dynamic energy analysis it is easier to go downhill quickly than uphill slowly. 

The foregoing arguments suggest that the profitability of nuclear power programmes in energy terms is far from 
established. Even for the best type of nuclear plant, the percentage of energy output required for investment 
is high — much higher than for equivalent non-nuclear power stations. 

Dr Chapman has calculated that the ratio P /P for a modern coal-fired power station is about 8, compared 
to about 3.3 for the best nuclear power plant. In a programme of coal-fired plant designed to achieve a doubling 
of output in T = 5 years, the percentage of output required for investment would be about 14%, compared with 
about 35* for the best nuclear plant in an equivalent programme.* 



4,2, CRITIQUE OF METHODS AND DATA 



In S3. 3, error estimates were assigned to power and energy ratios, with the caveat that the error estimates 
take no account of possible omissions or methodological simplifications in the analysis; the discussion in §4.1 
likewise assumed that the calculated ratios are correct. This section identifies and assesses these further 
sources of error which, if taken into account, may significantly change the results shown in §3.3 and §4.1. 

The foregoing analysis has bean deliberately biased in favour of high net energy yields from nuclear power 
programmes: technical uncertainties have been rather consistently resolved in such a way as to give nuclear 
power the benefit of the doubt. Below are listed various technical and methodological changes which might be made 
in the course of a more detailed and sophisticated future analysis. (The categories shown are for convenience only 
and are not always exclusive.) 

4.2.1. INVESTMENT INPUTS 

1. The present analysis has neglected the energy investment needed to build electrical transmission and 
distribution facilities, and to build any special equipment (e.g. hydroelectric pumped storage schemes) that 
might be needed to compensate for the effects of nuclear programmes on the economics and security of the grid. 
The capital cost of transmission facilities is roughly comparable to that of an equivalent capacity of conventional 
generating plant, so the corresponding energy requirement is presumably large. Such an energy investment is of 
course also required for fossil-fuelled electrical networks, but not for decentralized energy systems (whether 
electrical or otherwise) ; and it is the comparison of large-scale electrical generation with decentralized systems 
that is often of policy interest. 



Because advocates of nuclear power have claimed that nuclear power will be the dominant energy source in future 
and will substitute extensively for fossil fuels, this paper uses a one-to-one conversion factor in comparing 
fossil-plus-electrical investment energy with electrical output energy — a point explored further in §5, tl(c). 
The aim here is to study nuclear power more as an energy source than as a mere means of generating electricity.- 

investment energy. 



433 



2. The present analysis has neglected the energy investment needed to build supporting facilities for the 
nuclear fuel cycle — enrichment plant, reprocessing plant, fuel conversion and fabrication plant, transport facili- 
ties, etc. In an expanding nuclear economy, substantial energy inputs are needed to build these components of 
the ancillary industry, often well in advance of obtaining the resulting electricity outputs. Moreover, some 
support facilities, though very costly in money and energy, have in the past operated at rather low capacity 
factors, thus making poor use of the energy invested in building them. 

3. No allowance has been made here for recovering heavy-water inventory, and perhaps components, at the 
end of reactor lifetime and reusing them in replacement plant. This is the only "unconservative" assumption the 
author has identified, and its effect (especially at small T ) is very minor. Indeed, the effect of recovery and 
reuse can be seen by supposing that all the investment inputs of replacement reactors can be supplied by recovery 
from retired plant. The only effect of this assumption would be to delete the black areas in Figures 10, 11, and 
12 (page 9) . The percentages of output to investment and the payback times calculated earlier would be unaffected 
because they are already based on an assumption of infinite plant lifetime. 

4.2.2. PROCESS INPUTS 

4. Fuel-cycle process inputs of electricity, e.g. for enrichment, have been assumed not to entail inefficient 
conversion of original fossil fuels to electricity, but have been subtracted directly from reactor output. In 
contrast, much of the present global enrichment capacity, including most of that in the UK and USA, is in fact 
fuelled by conventional fossil-fuelled power stations. The efficiency of producing electricity by such methods 

is about four times less than is assumed here. 

5. Special process inputs to the fuel cycle have been neglected. They include production of hafnium-free 
zirconium, Magnox alloy, control-rod materials, nuclear-grade graphite, etc; some significant special process 
inputs have probably not yet even been identified. In general, exotic materials may have disproportionately large 
energy requirements, whilst unusual components made of common materials (e.g. large steel pressure-vessels) may 
have disproportionately low energy requirements. (In the UK economy, the energy requirement per £ value may vary 
as much as 30O-fold over the entire range of goods and services.) The fuel cycles treated here have been much 
simplified, and it is likely that some significant energy inputs to them have been missed, particularly for 
Magnox and HTR systems. 

6. Energy expended in safeguarding fissionable materials against theft or malicious release has been neglected 
in the present analysis, largely because the data are not available. For the same reason, maintenance energy has 
been neglected. 

7. The grade of uranium ore assumed here (0.3%) is larger than the average grade of ore typically being mined 
today (about 0.2%); the assumed grade yields unrealistically low process and capital inputs for mining and milling. 

8. The enrichment energy per unit output assumed here is somewhat smaller than in most present gas-diffusion 
technology, but probably larger than that which might be achieved by proposed gas-centrifuge or laser enrichment 
technology. Process savings through the use of gas centrifuges would be partly offset by their increased investment 
inputs (perhaps further aggravated if the design lifetime proves overoptimistic) , but the residual advantage might 
slightly reduce the energy requirements calculated here for enrichment. 

9. Plutonium extracted in fuel reprocessing and recycled into new fuel for burner (non-breeder) reactors has 
not been counted here as an energy credit against enrichment energy for the new fuel. Such a credit, which would 
be partly offset by increased process and investment requirements for handling the more toxic plutonium fuel, 
would be appropriate if and when plutonium recycle was begun, and might decrease total fuel-cycle process inputs 
by several percent if it became the nearly-universal practice. 

10. Process inputs have in general been calculated here on the basis of electricity sent out, not of gross 
electricity generated; hence process inputs needed to operate pumps, controls, active waste-heat-dissipating 
devices, etc at the power station could have been neglected here, resulting in underestimates of the required 
annual fuel inputs. 

4.2.3. OTHER INPUTS 

11. The present analysis neglects all energy used in research and development, both past and ongoing. This 
energy, if it could be evaluated despite difficult boundary questions, ought probably to be counted mainly as an 
investment input, as most of it was supplied before much nuclear power was produced. Another method of treating 
this input might be as a process input amortized over the lifetimes of all reactors built using the technology 
before it becomes obsolete; but that method is not so relevant to the societal question whether we should have been 
better off (in energy terms) not having a nuclear power programme from the beginning. 

12. Administrative overheads — design, safety analyses and precautions, other regulatory efforts, health-physics 
monitoring, accountancy, etc — have been mainly neglected here, especially those arising not in the CEGB's offices 
but in central bodies such as the UKAEA, Nuclear Installations Inspectorate, Department of Energy, Treasury, etc. 



434 



13. The methodology used here does not consider land-use as incurring an energy penalty. Some authors, such 

27 
as Odum , would disagree, arguing that commitments of land for a purpose which excludes settlement or some other 

productive uses displaces those uses onto other land — generally agricultural land — and thus increases the energy 

inputs needed to maintain the same level of productivity from the remaining pool of unaffected agricultural land. 

Odum would also argue that net-energy analyses should include contingent energy inputs arising from major 

accidents which contaminate large areas of farmland and thus entail large energy inputs to make up the lost 

biological productivity. The author of this paper, uncertain whether such a penalty is appropriate and lacking 

the data to calculate it, has omitted it. 

14. Possible energy penalties resulting from accidents, e.g. energy requirements for decontamination, evacua- 
tion, new construction, etc, have been neglected here: they might be considered an energy equivalent of insurance. 

15. Energy inputs associated with decommissioning defunct reactors have been neglected. 

16. Energy inputs associated with transport, treatment, storage, and disposal of low- and medium-level radio- 
active wastes (particularly noble gases and low-level alpha emitters such as cladding hulls) have been neglected. 

17. All process and investment requirements for transporting, treating, storing, retrieving, safeguarding, 

and disposing of high-level wastes have likewise been neglected. The disposal technologies are still speculative, 

2 
as is the appropriateness of the storage options proposed . In the absence of credible disposal methods, one 

might calculate the energy which would be required for long-term storage if it were to rely on the methods that 

are now proposed to be used for interim storage (for periods ranging from a few decades to a century): over the 

very long periods required, these energy inputs would be comparable to, or would exceed by as much as about an 

hundredfold, the lifetime output of the reactors served. The omission of all "future services" inputs from the 

present analysis may thus be significant even in the static case of one reactor in isolation. 

4.2.4. OUTPUTS 

18. The present analysis has assumed that the lifetime average capacity factor of nuclear power stations is 
62% — in effect, that the energy invested in them is used with 62% efficiency to produce useful outputs. However, 
the lifetime average capacity factor assumed for forecasting by the US Atomic Energy Commission is 57.375%; 
and recent USAEC data suggest that the actual value may be substantially lower, say about 50%, since aging 
problems (metal fatigue, corrosion, accumulation of dirt, etc) are more serious and cure appearing earlier than 
expected. The present results are sensitive to the capacity factor assumed. 

19. All nuclear power stations have been assumed to operate throughout their lifetimes at their full design 
power rating. In contrast, very many reactors in various countries have been derated owing to corrosion, new and 
less sanguine safety analyses, etc, and there is no guarantee that this will never happen again. Derating, like 
reduced capacity factor, has a direct effect on the energy ratio. 

20. No account has been taken here of possible delays in construction and commissioning, nor of prolonged 
shutdowns not considered in the average capacity factor (e.g. those due to strikes), nor of the possibility that 
all or part of the nuclear industry might be shut down in the event of a major accident or safeguards failure 

or of certain political developments at home or abroad. 

21. The values assumed for construction time and delay time are substantially shorter them those actually 
experienced in most countries, and thus probably give rise to unrealistically favourable dynamic energy yields 

(§2.3, especially Figures 16 and 17 on page 11). Indeed, in many cases the construction and delay times are 

26 
observed to be steadily increasing . . 

22. Advocates of nuclear power often postulate "learning curves" whereby increasing experience leads inevitably 
to improved reliability and quicker construction. Examples of such behaviour in other technologies can be adducedj 
but in the case of nuclear power, the "learning curve" hypothesis is inconsistent with much experience ' and 
must rest on unique and well-documented causal relations, not on casual analogy. This is especially true since 

the doubling time of reactors far exceeds the doubling time of the numbers and skills of the most competent and 
dedicated members of the nuclear community. 

4.2.5. GENERAL METHODOLOGY 

23. The present analysis does not give credit for large-scale use of waste heat from nuclear power stations. 
Near-urban siting to facilitate district heating seems neither wise nor likely, and major uses for industrial 
process heat will probably be constrained by problems of reactor reliability (as failure of process heat can be 
very damaging to industrial equipment relying on it) . If these difficulties can be overcome, the assumed energy 
outputs should of course be increased accordingly. 

24. Contrary to experience, the present analysis has assumed that altered or additional supporting processes 
and facilities will not be required in future nuclear fuel cycles. The history of the nuclear industry shows that 
costly (in money and energy) new devices and processes are often required as more is learnt and as public contro- 
versy persists . (It is also possible that future technical developments might decrease energy inputs to nuclear 
power, though it is not clear where, save in enrichment — q.v. 11118,9 above,) 



435 



25. The present analysis has assumed throughout that the energy required to produce a unit of each material 
(steel, copper, cement, etc) will remain constant in time. In practice, some energy requirements are likely to 
decrease through technical innovation whilst others increase through declining ore grade, geopolitical constraints, 
or the like. 

26. Likewise, the energy requirements of producing energy inputs to the nuclear fuel cycle have been assumed 
to be constant. In view of present experience of increasingly marginal fossil-fuel technologies, this assumption 
is probably overoptimistic even on a time-scale of one to three decades; in the long run it is certainly too 
sanguine, owing to the requirements of the Second Law of Thermodynamics. More specifically, in the electricity- 
generating sector no allowance has been made for either existing fossil-fuelled plant or existing reactors being 
demoted in the merit order at any particular time (probably reducing both system efficiency and the capacity 
factor of older nuclear plant) as many new nuclear stations are commissioned. 

27. The present assumption that reactor outputs are uniformly distributed over the lifetime and reactor 
inputs over the construction time may be unrealistic, though in which direction is unclear. It appears ' that 
outputs are likely to decrease during most of the reactor lifetime (after an initial increase during the 

first few years, owing to the resolution of "teething troubles") whilst process inputs gradually increase owing 
to increasing maintenance needs and to the increasing energy requirements per unit of energy and materials inputs. 

28. The present study has presented analytic solutions rather than detailed (unsmoothed) simulations, and 
has not been able to deal with a mix of reactor types, such as possible mixes of burner and breeder reactors, 

or of different types or designs of burner reactors, or of reactors using thorium fuel cycles. Thus the analysis 
is indicative rather than definitive, and is not a substitute for detailed simulation in any particular case. 

4.2.6. SUMMARY 

A more detailed treatment of these matters, with the possible exception of 118, 9, 18, 22, 23, 24, 25, and 27, 

would probably lead to net-energy calculations substantially less favourable than those presented in §§3, 4. 

The author believes that the energy ratios shown in §3.3 are upper limits and are likely to be significantly 

reduced by further analysis, thus leading to even worse results in 54.1. As will be readily appreciated, however, 

2 
it is probably not possible to quantify some of the important terms . 

5, SOME QUESTIONS AND ANSWERS 



recurring questions which seem worth answering here. 

1. If this analysis of nuclear power programmes is correct in showing that they can be net consumers of 
energy, however can they appear economically attractive? Why does energy analysis give a different answer 
than economic analysis? 

This question, with variations, is the most common reaction to the analysis. Four points need to be made 
in reconciling the apparent conflict between energy and economic analyses. 

a) The market only reflects people's perceptions. If people like those who need to ask the question do not 
perceive the dynamic net-energy problems of nuclear programmes, the market will not do so either. 

b) The economics of nuclear power are not as clearly favourable as some pundits would make out. Most eco- 
nomic analyses of nuclear power have been done in order to promote nuclear technology. This bias is reflected in 
omitting to take into account such items as the real capital costs of nuclear plant , the costs of future 

services (54. 2. 3, 117, and 110 below), the capital costs of support facilities, the increase in uranium costs 

26 
owing to the increase in conventional fuel costs, and the capacity factors likely to be attained in practice . 

It may be correct that when costed properly, single nuclear power stations may be preferable to single fossil- 
fuelled power stations; nobody really knows. If the need to offset a substantial fraction of the output of a 
nuclear programme against investment in it were treated as a cost overrun for the constituent power stations, 
they would probably look most unattractive. 

c) The comparison made in economic analyses of nuclear power is between nuclear-generated electricity and 
electricity generated from fossil fuels. It is often said that if nuclear power were a net energy consumer, it 
could not compete with other sources of electricity. But look at those sources! Nobody would suggest building 
oil-fired power stations in order to relieve an oil shortage, for everyone knows that fossil-fuelled power stations 
are net consumers of energy, consuming between three and four kWh for every kWh of electricity sent out. So if a 
nuclear station consumed, say, two kWh of fossil fuel per kWh of electricity sent out, but if the energy content 

of the uranium also consumed were taken to be gratis, then it would appear, in energy terms, that nuclear fission 
is a preferable technology for generating electricity. Yet under these conditions neither fossil-fuelled stations 
nor nuclear stations would be considered a good method of Increasing the world's fuel resources. (To show, by 



436 



- 21 - 

economics, that nuclear power is a net producer of energy would require a demonstration that the total cost par 
energy unit from a nuclear power station is less than the total cost per energy unit from a coal-mine or oil-well.) 
Let us compare two systems for generating electricity: one from fossil fuel alone, the other from uranium plus a 
fossil-fuel subsidy. The latter system stretches the fossil-fuel reserves and thus is a more resource-efficient 
way of using them to generate electricity ; but that does not mean it is a good source of energy. 

d) The final point concerns the objectives of energy analysis. It is not a system of allocating resources; 
it is not a replacement for economics. All that energy analysis can do is to show the energy implications of 
policies. In contrast, economic analysis shows the financial implications of policies. These two approaches are 
not contradictory but complementary. The mistake implicit in the question is the assumption that a fuel industry 
which makes a fiscal profit must be a net producer of energy. The CEGB is a good example of why this assumption is 
wrong. Energy is just one of the resources that economics should consider, and can only be considered the most 
important one if the policy being analyzed demands that energy be so treated. 

Thus, in summary, the answer to the question is that nuclear power can show a fossil-fuel profit when compared 
to a fossil-fuelled industry which is a prolific net consumer of energy, but this does not imply that nuclear power 
is a net producer of energy. 

2. It is not clear why this energy analysis is necessary, since if you are right, the price of uranium and 
of nuclear-generated electricity will rise, giving us due warning that we are heading for trouble. 

See answer (a) to question 1. Moreover, if you wait until the somewhat ponderous price mechanism has got 
round to reflecting the actual situation of net energy consumption — i.e. difficulties perceived as present or 
imminent — you are already so far into a net energy deficit that you are in real trouble. The market is notorious 
both for long perceptual delays and for vigorously signalling short-term abundance of various resources when what 
matters for policy is the value corresponding to long-term scarcity of the residual stocks. 

3. If your analysis is correct, what should be done? 

This analysis shows that building large numbers of nuclear reactors does not increase energy supplies; so do 
not build them. 

To show that a policy is mistaken does not oblige the analyst to have an alternative policy. As it happens, 
however, the author and his colleagues do have ideas explored in past ' and forthcoming papers from Friends of the 
Earth, and hinted at in §6 of this paper. According to these ideas, the human and fiscal resources now devoted to 
nuclear power can and should be redirected to other energy supply and conservation technologies with more favourable 
energetic and social characteristics. 

4. Your analysis has confused energy with welfare, and it is not obvious that this type of analysis is at all 
relevant to nuclear policy. 

This comment was not invented; it is an accurate summary of one received from the Department of Energy. The 
reply was that the report referred to nowhere confused energy with welfare, and that the only justification for 
looking at the energy implications of nuclear power programmes is that it is claimed (by their proponents) that 
they increase the supply of energy. If nuclear power stations are built for other reasons, then energy analysis 
may well be irrelevant to the formulation of nuclear policy. 

5. You have only demonstrated that nuclear systems are net consumers of energy under ridiculously high rates 
of growth. If you used a more sensible growth-rate, your conclusions would be more favourable. 

Precisely. The UK nuclear growth-rate assumed, however, is that officially projected by the CEGB, UKAEA, 
and similar authorities, all of whom are working hard to expand nuclear capacity with a doubling time of about 4>j 
years for at least the next 20 years. This growth-rate is not that projected by the same authorities for total 
energy demand; it arises because nuclear power is being put forward as a source able to take over the role of 
oil and other fossil fuels even as their availability may decrease. The results are indeed ridiculous, but are a 
straightforward extrapolation of historical trends . (See also Figures 1 and 2.) It is noteworthy that the 
nuclear growth-rate officially projected for the UK is among the lowest in the worldi the approximate doubling 
times projected are about 3 years for the EEC, 2S for the USA, and 2 for France. 

6. Even if you have got the correct rate of growth, nobody is pretending that it will go on indefinitely , 
since breeder reactors will take over the growth within 15-20 years. 

This analysis does not pretend that growth will continue indefinitely at the projected rate or at any other 
rate. This paper considers only a transient problem whilst the nuclear burner programme is expanding rapidly, 
and shows only that burner reactors (relying on high-grade uranium resources) are net consumers of energy during 
this period of rapid growth. Their viability in future will depend on uranium ore-grade, changes in nuclear 
technology, and other variables. (See also, however, the discussion in 56 on the recovery of cumulative energy 
debts after growth has moderated.) 



437 



If breeder reactors are to represent a large part of the new nuclear capacity installed, it will be necessary 
to continue to increase the number of burner reactors of the types considered in this paper. This is because a 
breeder reactor takes about 15-20 years in likely future practice (nearer 40 in present practice) to breed enough 
Plutonium for the core of a second breeder reactor, and longer still (by as much as about twofold) to breed the 
"pipeline inventory" required for the fuel cycle of the second breeder. Breeder reactors by themselves are thus 
constrained to a doubling time of some decades. To build up a stock of breeders at a faster rate would require 
some source of plutonium other than breeders. The only other sources of plutonium are weapons stockpiles and the 
Plutonium output of burner reactors. Fast growth in breeder reactors thus requires either the depletion of weapons 
stockpiles (which are too small to suffice for long) or earlier fast growth of burner reactors — coupled with 
concurrent operation of the breeder and burner programmes for a very long time (typically about a century) . The 
problem of plutonium stocks and flows in a breeder economy is closely analogous to the flows of energy in programmes 
of nuclear power stations: just as the breeders, if built at an excessive rate, can no longer breed (collectively, 
not individually), so the nuclear stations, if built at an excessive rate, cannot provide their own energy investment. 

Several more detailed comments and questions are also worth addressing here: 

7. Why is the initial fuel charge counted as an investment input? 

Because a nuclear power station cannot operate at all until the entire core has been assembled. This is a 
true investment: all the energy (and money) for the initial core must be paid out before any energy output can be 
obtained. The initial core actually represents about an eighth of all the fuel that the reactor requires over its 
25-year lifetime. For a 10OO-f«(e) coal-fired power station, this would be equivalent to a situation where of the 
order of 14 million tonnes of coal had to be mined and delivered before the station could be started up. 

8. Since the core is intact at the end of the reactor lifetime, it should be included as an energy credit 
in the total output of the reactor. 

If a reactor core is still intact when the reactor is decommissioned, then it could change the energy ratios 
for second-generation reactors. It has absolutely no effect on the transient problem analyzed in the paper: 
apart from anything else, it should be clear that in a growth period, the number of reactors being started up is 
always larger (and generally much larger) than the number being closed down. (See also §4.2.1, H3.) 

It is also doubtful that anyone would continue to refuel a reactor with normal charges right up to the point 
of shutdown. It seems more likely that when a reactor is decommissioned, the recoverable fuel values from the 
core will be substantially lower than those from an equivalent amount of fuel during the normal operating lifetime. 

9. You have not given any credit for the plutonium produced. 

a) If you count the energy value of the plutonium as a credit, then to be consistent you should count the 
energy content of the uranium fuel as a debit, on the same principle used for fossil-fuelled stations. For both 
uranium and plutonium, the raw material is of no use to anyone until it has been converted into a useful form. 
This analysis is concerned with the energy profitability of energy conversion processes, not with the potential 
energy of the raw material. 

In terms of real energy flow into an economy, the plutonium is irrelevant unless it is actually recycled 
(§4.2.2, t9) . This paper has shown that a rapidly growing population of burner reactors consumes more energy than 
it produces. It is not much use offering plutonium to people who want electricity. Thus plutonium is not relevant 
to the transient problem that concerns us here. 

b) The plutonium is not normally considered a fuel for burner reactors (though it could be: §4.2.2, 19). 

It also cannot be used in a large-scale breeder programme for a long time. If plutonium production were credited 
to burner reactors because it was actually being recycled and did actually save enrichment energy, or likewise for 
breeder reactors, the same plutonium could not be double-counted as a further credit in the systems thus served. 

10. You have not included any energy used in waste handling, storage, or disposal. 

This is because no realistic data on these operations are available. Any energy expenditure on these opera- 
tions would only make the situation worse than portrayed here. In the long term, these energy inputs may be 
important, though they may have little immediate relevance to the transient problem analyzed here. 

The most abundant transuranic waste material is plutonium-239 with a half-life of 24,390 years. Although 

plutonium is recovered from spent fuel rods, 100% recovery is not economic or technically practicable (and is 

probably not energy-profitable either) . Thus waste from spent rods contains some residual plutonium and must be 

2 
stored at least, say, 240,000 years (10 half-lives) — probably a very conservative estimate . Now if a power 

station produces lOOO MW for 25 years but leaves waste materials which require a power input for maintenance of 

lOO kW for 250,000 years, then the net energy output (neglecting all other inputs) will be zero. A power input 

of 100 kW is equivalent to an annual consumption of 90 tonnes of steel or 25 tonnes of rolled stainless steel. 

One hopes that designers of waste-management schemes will be able to keep their average power consumption below 



438 



100 kW per reactor as well 

It may be of interest that the waste-storage methods now proposed in the USA (pending development of a viable 
method of permanent disposal) entail an average power consumption of about 12-35 kW„ Specifically, if Chapman's 
data for gross energy requirements of various materials are applied to the data in the USAEC's draft Environmental 
Statement on the proposed Retrievable Surface Storage Facility , then the continuous power input to build and 
operate such a facility (assuming the lOO-year design life stated) is found to be about 3—9 MW(t), depending on 
the design chosen. The facility will serve approximately 10 reactor-years of PWR waste production, corresponding 
(at 1 GW(e) /reactor and 62% capacity factor) to a total lifetime output of the order of 6 x 10 MW(e) -years. 
Thus the total energy inputs to a series of successive facilities of this type would be of the same order of 



(corresponding to the ZJ3 Pu decay chain) and would be 'vlOOx larger than the total outputs if the storage time were 
about 10 years (corresponding to the 2 '* 1 Am- 237 Np decay chain ). Of course, less energy-intensive means of safe 
storage or disposal might be devised and demonstrated, but none appears to be available now for engineering assess- 
ment, and proposed methods for partitioning and recycling actinides may not be of much help . Moreover, throughout 
this paper it has been assumed that the energy requirement for supplying the energy inputs is constant in time; 
whereas the Second Law of Thermodynamics entails that a joule of free energy is easier to earn today than tomorrow 
(i.e. that energy has a small negative discount rate) — a conclusion fully borne out by recent experience with 
marginal energy technologies. This conservatism of the above calculation should become quite important over the 
very long periods relevant to waste management. 

11. You have counted electricity consumed in enriching initial fuel charges at the thermal rate based on 
conventional fossil-fuelled stations. However, as each lOOO-MW(e) reactor is built, the average efficiency of 
generating electricity will steadily improve, thus making the energy requirements of reactors steadily decrease. 

This comment is invalid because whilst the nuclear power programme is running at a net energy loss the 
efficiency of generating electricity is actually declining. Thus by maintaining one efficiency throughout the 
growth period our calculation has been overgenerous. If ever a nuclear programme did move into net energy profit, 
then it would be correct to use an improving efficiency of generating electricity. 

12. You have only considered gas-diffusion enriclment of uranium. If you had considered gas-centrifuge 
enrichment, which requires a tenth of the power input, your conclusions would be changed. 

No reliable operating data seem to be available to confirm the estimated requirements of gas centrifuges. 
Even if the centrifuge has a tenth of the present operating power, its greater capital intensity may partly 
compensate (§4.2.2, 118). At the moment the enrichment process is the most energy-intensive step in uranium 
processing. The investment inputs to gas-diffusion plant contribute about 10% to the total energy inputs to 
enrichment; the centrifuge process has been estimated to be more capital-intensive and to have some higher 
operating costs. Hence the energy requirement for centrifugation is likely to be substantially more than 10% 
of the energy requirement assumed here for gas diffusion. Moreover, though use of a gas centrifuge would 
increase the energy ratio for any one reactor, the shift of energy requirements from process inputs to investment 
inputs would make the dynamic problem worse. Thus until better data are available, one cannot be sure whether 
or not the centrifuge process will improve the dynamic energy flows at all, let alone significantly. 

13. You have included uranium mining in your energy analysis, even though the UK does not mine any uranium. 
Surely this makes nuclear power very attractive as a net energy source for the UK. 

The energy requirements for mining uranium are an insignificant terra (affecting the energy ratios by only 
about 0.2, well within the error estimates) if high-grade ores are assumed. The important energy requirements, 
fuel enrichment and station construction, are supplied by the UK. In future, as the ore-grade declines, the 
energy requirements for mining will increase until, at grades less than about 0.01%, they will dominate the 
analysis. At this point one must ask whether it is reasonable to expect uranium exporters to subsidize one's own 
fuel-production system. This is essentially a political question: energy analysis can only show the energetic 
implications, it is unpalatable to have to decide whether dependence on uranium exporters is preferable to 
dependence on OPEC. 

6. CONCLUSIONS AND POLICY IMPLICATIONS 

-i.- ^aper presents a conservative assessment of the net energy yields from various thermal reactors, neglec- 
ting a number of terms whose inclusion would be likely to make the results worse by a significant but presently 
(and to some extent perhaps perpetually) unknown amount. This optimistic assessment shows that reactors fuelled by 
uranium from high-grade ores typically yield output energy about two or three times as quickly during their life- 
time as they consume input energy during their construction. If the uranium is derived from low-grade ores (in 



439 



particular, Chattanooga Shale) , this power ratio becomes about h — l 1 *, and an isolated thermal reactor produces 
(in most cases) rather little excess energy beyond that required to fuel it. 

In the dynamic rather than the static case — that is, if reactors are considered not singly but as part of a 
multi-reactor nuclear power programme whose total inputs and outputs at various times are of interest — the 
requirements on the output-to-input power ratio of the individual reactors are far more stringent. If the number 
of reactors increases too quickly, the energy /year that must be continuously invested in new construction is a large 
fraction of the programme's output. For example, if a nuclear power programme based on the most energy-profitable 
type of reactor studied (assuming high-grade uranium ores) is to yield to society half of the energy/year that it 
was expected to yield (the other half being offset against investment in the programme) , then the doubling time 
of the number of reactors cannot exceed 4.0 years; the corresponding figure for the least energy-profitable reactor 
studied is 5.5 years. And if the doubling time of reactor population exceeds 2.6 years for the most energy- 
profitable reactor, or 3.5 years for the least, then the nuclear power programme will continuously consume more 
energy/year than it produces. For comparison, the doubling time widely proposed for the British nuclear programme 
is about 4.3 years (Figure 6); for the EEC, about 3 years; for the USA, about 2.5 years; and for France, about 
2 years. All these programmes will therefore produce, as output to society, energy/year equivalent to less than 
about half the demand they were intended to meet, and the output of at least the last two programmes (the US and 
French) will be negative. 

The shaded area in Figure 15 (page 11) shows the constraints imposed by nuclear doubling times that are 
short enough to achieve a significant measure of substitution for oil within several decades. The doubling times 
included in the shaded area — from 2.0 to 4.5 years — cover the range of historical and officially projected nuclear 
growth in nearly all countries. All these doubling times are short enough to ensure, with attainable power 
ratios (<3) , that no more than about half the expected output is available to society for purposes other than invest- 
ment in new nuclear plant, and that in most cases this available energy is far smaller or even negative (correspon- 
ding to continual subsidy from fossil fuels to nuclear power). 

In the British case, with reactor population proposed to double every 4.3 years or so, and assuming high-grade 
uranium ores, only about a quarter to a half of the energy which the programme is supposed to produce would actually 
be left over for general use after reinvestment in the programme. This can be interpreted as meaning that to meet 
a given final demand by society in general (excluding the nuclear industry) , about two to four times as much capacity 
must be built as was expected; alternatively, that a unit of net output to society from the nuclear programme will 
cost two to four times as much as had been claimed. With uranium from low-grade ores (Chattanooga Shale) , a sus- 
tained programme of thermal reactors with a 4.3-year doubling time would always be a net consumer of energy: the 
more reactors we built, the more energy we should lose (see Figure 15). 

Two types of policy conclusions flow from these considerations. The first is that doubling times of a few 

years (2-3 and probably 4) are certainly not sustainable with nuclear power if it is to produce net yields; indeed, 

this is probably true of all other major new energy technologies. Such rapid doubling times are the result of 

trying to meet assumed rapid growth in demand whilst simultaneously substituting rapidly for traditional sources 

(such as Middle Eastern oil). Major new energy technologies can, for a time, do one or the other but not both; or 

3 
at least if they do both they cannot also produce net energy. This is a good argument (if one more were needed ) 

for stabilizing or reducing energy demand if one wishes to buy the time needed to substitute — a need many people 

had already perceived owing to the intractable rate-and-magnitude problems of a purely logistical nature that 

impede very rapid proliferation of any complex technology. 

This argument has special force in countries like Britain which are contemplating several kinds of substitution 
at once: North Sea oil for Middle Eastern oil, synthetic fuels for original coal, nuclear power for oil and coal, 
etc. These substitutions, too, have their net-energy implications and rate constraints, entangled at the root 
with added demands for money, materials, technical skills, and other scarce resources. 

Another set of policy conclusions is suggested by the question: When are more favourable cumulative net yields 
of net energy obtained if an exponential programme, initially too rapid, is later moderated or terminated? No 
exponential process can continue forever; but what are the energetic consequences of reductions in growth-rate? 
The answer depends in detail on the conditions assumed, but in general the time taken to recover cumulative energy 
deficits incurred by too-rapid initial growth is very long; it may exceed the lifetime of the energy facilities being 
built. Moreover, it appears to be true that at least in the early stages of a too-rapid programme that Is incurring 
a deficit, it is less painful in the long run (and a more efficient use of national resources) to stop the programme 
and promptly redirect its resources in a more rewarding direction than merely to slow the programme in the hope of 
recouping losses incurred so far. (This conclusion is tentative: much more work is needed In this area.) 

An analogy from economics may be helpful here. If an industrial economy with a fixed rate of capital formation 
grows exponentially at an excessive rate, its rate of internal capital formation will be too slow to support invest- 
ment, so an exponentially increasing external subsidy (bearing a fixed proportion to output) must be obtained. If 



440 



growth-rates are later reduced, or if the profit margin available to repay the external debt is increased (an 
analogue of improving the power ratio of energy-conversion facilities) , then one may be able to obtain enough sur- 
plus money to pay back the debt over some period. But if the surplus income disposable in this way is not very 
large, the time required to repay the external debt may be very large; and if it exceeds the lifetime of the plants 
being built with the borrowed capital, there is a sense in which it would have been better not to build them: they 
contribute to a later profit, but are not themselves profitable during their lifetimes and do little to make one 
richer. The capital sunk in them can sink out of sight. (As noted in §5, 16, there is also a material-flow 
analogy with plutonium in a breeder-reactor economy: consider the case where the time required to repay the plutonium 
debt exceeds the reactor lifetime.) 

Persons familiar with cash-flows and with profit-and-loss accounts will readily appreciate that accumulating a 
large external debt in the way just described is not sound management, and should at least be confined to as short 
a period and as small a sum as possible. (Life is harder in the financial world, where one must pay interest at a 
high rate; but as pointed out in §5, 110, there is a sort of interest rate on energy borrowings too, since one is 
borrowing high-grade energy of low energy cost — e.g. Middle Eastern oil — but will repay the debt with lower-grade 
energy of higher energy cost — e.g. North Sea oil or an even more marginal source. If one waits too long to repay 
the energy debt, the interest payable on it may be truly exorbitant and exceed the principal sum — e.g. the grade of 
uranium ore available may be very low, reducing P /P. — so that the belated energy yields may no longer suffice 
to repay the debt.) 

Unfortunately, managers who choose to retard or terminate excessively rapid exponential growth in, say, 
nuclear power may create serious new problems in their effort to solve present ones. For example, whatever rapid 
construction rate is chosen now must be essentially duplicated with an "echoing" tranche of new capacity about 
25 years later (or whatever the reactor lifetime may be) , since reactors built within a short period will all have 
to be retired, and presumably replaced, within a similarly short period. If growth has been slow after the initial 
crash programme, that "slump", too, will be "echoed" 25 years later. The result will be a violent oscillation of 
construction rate, not conducive to the smooth functioning of the national fiscal or industrial machinery; the 
oscillation can be gradually damped over several generations of reactors by retiring some before and others after 
their nominal lifetimes have expired (thus decreasing the energy ratio of the former and perhaps increasing that of 
the latter) , but meanwhile the oscillation will be very disruptive to the nuclear industry and to economic functions 
related to it. Clearly the best way to minimize this sort of transient problem is to abandon crash programmes as 
quickly as possible: the longer one waits, the worse the transients will be and the graver their social and economic 
implications? Once a country is committed to nuclear growth that is rapid enough to run at an energy deficit, 
either the growth must continue (thus further enlarging the energy deficit) or the growth must slow down (thus pro- 
ducing embarrassing transients for some decades thereafter). One obvious way to avoid both problems is not to under- 
take the growth; urgently needed further research may reveal other and more complex strategies. 

One desirable transient, however, is that as soon as investment in new plant is reduced or stopped, a tranche 
of new supply of energy (and money, materials, skills, etc) becomes immediately available for use in other 
energy supply or conservation efforts, or for other social purposes. It is often said that energy supply from e.g. 
North Sea oil and nuclear power should grow very quickly now so as to create capital for investment in other 
sources of supply, e.g. direct and indirect solar. Yet the sudden release of resources (previously committed to 
nuclear power) when rapid investment in nuclear plant abates, though it demands careful planning to minimize dis- 
ruption, may be a readier and, in the long run, a less disruptive source of the energy and other resources needed 
to realize the long-term goal of harnessing energy income (i„e. renewable and inexhaustible energy sources). 

If nuclear power were a "soft" technology of smaller scale and smaller energy inputs, it would be less painful 
to reduce rapidly our investment in it, and we should have more room for mistakes, more leeway for exploration. 
But there is a connexion between this exploration and our nuclear policy: the British institution best suited to 
apply its undoubted talents to "unconventional" energy options and to energy conservation is probably the 



nuclear research and development establishment. This impressive technical resource could be usefully redirected 
to less centralized technologies with more favourable energy dynamics. 

Traditionally it has been supposed that our energy problems can be solved only by substitution of fuels: 
wood, coal, oil, gas, uranium, — plutonium? But viewed from the broader perspective of modern strategic thinking 
about energy, this seems a rather short-sighted view that has arisen only recently through a series of historical 
accidents. Sophisticated societies in many parts of the world have worked for a very long time without large 
injections of fossil or fissile fuels, whose widespread use has lasted only a few centuries and whose massive use 
has been a very short-term phenomenon of the past few decades. Today's energy problems arise not from shortages 
of energy but from shortages of our specifically habitual drugs of addiction — and from shortages of vision. Later 
papers in the current Friends of the Earth series on energy policy will explore in detail the problems and prospects 
of various sources of energy capital and energy income, and of their utility in achieving particular social goals. 



•Implicit in this discussion is that short- and medium-term energy stringency means that whatever one's strategic 
goals, large amounts of immediate investment energy are simply not available. 



441 



brushed aside with the contention that whatever its failings, nuclear power is at least an abundant long-term 
source of energy with which to cure our social ills, If, as the analysis in this paper suggests, nuclear power is 
"an all but infinite source of relatively cheap and clean energy" , and under certain circumstances {e c g. deploy- 



ment fast enough to seem a useful substitute for oil) is in fact an energy sink, then these familiar arguments 
for nuclear power must be re-examined. 

The pioneering analysis by Chapman and Mortimer has, like all first explorations of a new field, raised more 
questions than it has answered; so too has this paper, which the author hopes has gone further to clarify basic 
principles of dynamic energy analysis. (Still further clarification at a greater level of complexity will be 
provided in a forthcoming joint technical paper.) Where this paper has modified the data or approach of Chapman 
and Mortimer , it is not to denigrate their work, but to refine and enlarge it in the way that they themselves 
would wish and (in many cases) have directly suggested. Further refinement of methodologies and data is clearly 
needed for this paper as much as for theirs: and if this paper does nothing else, it should demonstrate the 
existence of legitimate and important net-energy questions whose prompt and precise resolution is important 
to policy decisions now being taken throughout the world. It is long past time to resolve the ambiguity latent 
in the Institute of Fuel's recent conclusion that "Nuclear power generation must take over a large part of the 
increase in energy demand by 1980. " 



APPENDIX: EXPONENTIAL GROWTH 
AN ANALYTIC TREATMENT OF ENERGY CONVERSION PROGRAMMES DESIGNED TO SATISFY EXPONENTIAL DEMAND 

If P is the annual output of energy (net of process inputs) from an energy conversion plant and n,(T) is 
the number of such plants operating at time T, then the energy output/year at time T is E (t) = P n,(T). 
Consider the case where the number of operating plants is given by 

n f (T) = N o e a(T " T c " T p> (A.l) 

where N is the number of plants operating at time T = T + T 

T is the time taken to construct a single plant 

T is the time pause or delay between the end of construction and the start of operation 

a = (In 2) / T D 
where T is the exponential doubling time of the programme. For the moment we assume that the operational lifetime 
of the plants T is infinite. If n„ (T) plants are operating at time T, then the number under construction then is 

nC<T)=N ° e aT -aT ^^^ (A.2) 

n (T) = N e (1 - e c ) for T >_ T, 

The annual energy investment in construction of plants at time T is then 



X ± (T) = P i n <; (T) 



for T < T 
aT,. C (A« 3) 

P.N e (1 - e c ) for T > T 



i 

where P is the average annual energy investment per plant during the construction period. The energy output from 
operating plants at time T is 

X Q (T) = p o N e~ a(T c + T p'e aT for T > (T + T ) (A. 4) 

If the operative lifetime of the plants is T , then after T £ years the building programme must begin again to 
ensure that decommissioned plant is replaced. After each subsequent T £ years, replacement must again be taken into 
account. The expression for * L (T) (A. 3) is based on the assumption that the operational lifetime of the plants is 
infinite. For finite T £ it becomes 

X,(T) - P i N n eaT for < T <T i (A. 5) 

-aT Cl aT , „ . _ }first generation 



for T < T < (T + T ) 1 


first 1 


1 — I c 


1 second 


for (Tj •■ T c ) < T <_ 2T t j 


generation 



68-3'Jl O - 76 - 29 



442 



-aT, 



W 1 



) (1 
)(1 + 



-aTj 



-2aT 1}e aT 



for 2T Z < T <^ (2T 
for (2T. + T ) < 1 



,-aT^,, ^ ,-aTj + e -2aT i)e aT 

x 

(A. 5 continued) 

and so on. From A. 4 and A. 5 the investment energy/year as a fraction of output energy/year is 



T J^lst, 2nd, l 
3T }3rd genera- 
£•* tions 



C o (T) 


M 

R 

(1/R)(M + e" a(T * " T c" V) 
(M/R) (1 + e" aTl ) 






(1/R){M(1 + e- aT *) + e- a(2T ^ 


" T< 




(M/R) (1 ♦ e" aT * + e" 2aT ^ 





for (T 

for r l < 
for (T^ 
for 2T t 
for (2T„ 



r < (T^ + 

T e ) < T ; 

T <_ (2^ 
f T ) < T 



3T 



where M = 



1) and R = P /P. 



(Note that M is independent of time. ) So that within each time period 



the fraction of the energy output/year that is required for construction of new plant is also independent of time. 



This fraction increases with the replacement of plant. In Table A.l the values of e *■ , 
for a selection of exponential doubling times T and for both T = 25 and T = 50 years. 



-2aT x 



etc are listed 



T D (years) 




2 


4 




S 




10 




20 




T £ (years) 


25 


SO 


25 


50 


25 


50 


25 


50 


25 


50 


-aTj, 
-2aTi 












0.0133 
0.0002 
0. 


0.0002 

0. 

0. 


0.0313 
0.0010 
0. 


0.0010 

0. 

0. 


0.177O 

0.0314 
0.0055 


0.0314 
0.0010 
0.0002 


0.4200 
0.1764 
0.0740 


0.1764 
0.0311 
0.0055 



TABLE A.l. VALUES OF e 



r * FOR DIFFERENT DOUBLING TIMES AND FOR T = 25, 50 years. 



Analogous values are obtained for higher generations. (The table shows values less than 0.0001 as zeroes.) 

Clearly from Table A.l the effect of replacing obsolete plant is most important for programmes with a long T . 
Consider a programme with T = 5 years. If the plant lifetime is 25 years, then the fraction of annual energy output 
required for investment is 



a(T r 



M/R 

(1/R) (M + 0. 

1.0133M/R 

(1/R) (1.0133M + 0.0002e 

1.0135M/R 



a(T- + T_). 



for 


<T r 


for 


T * 


for 


(T, 


for 


2T £ 


for 


(2T 



2T * 

► T„) 



But for T = 20 and T = 25 the equivalent fractions are 

M/R 

(1/R) (M + 0.4200e a(Tc + T P') 

1.4200M/R 

(1/R) (1.4200M + 0.1764e a(T c + T P ) 

1.5964M/R 

and so on. If we now define P(T) as the annual energy output at time T after subtracting that needed for 
investment, then 



for 


<T C 


♦ v 


< T < 


- T £ 


for 


T £ < T < 


(T^ + 


V 


for 


(T £ 


♦ v 


< T < 


- 2T l 


for 


2T l 


< T < 


(2T^ 


+ v 


for 


(M 


+ T 


) < T 





- <W R)e 

- (P Q N /R) (1 - e" aT C)e aT 
(1 -°(M/R))P o N o e" a(T = + T P 


) aT 


{ (1 - (M/R))e~ 


a(T c + T p ) 


d/R)e" 



" aT MP N e a 
o 

{(1 - (M/RM1 + e- aT ^))e- a(T c + ^ S\* Q 
{(1 - (M/R)(l + e" aT *)e" a(Tc + V - 



tor 


< 


T 1 T c 




for 


T 

c 


: T <_ 


(T c + T p ) 


for 


(T c 


♦ v 


< T 


£T £ 


for 


T £ <T< 


w, ♦ v 


for 


(T * 


♦ v 


< T 


<2T £ 



(1 - (M/R) (1 



aT 
CO 6 
+ e - aT i + e - 2aT l 



>)P N _e 



-a(T„ 



T p ) e aT 



and so on. Figures 10, 11, and 12 are plots of X (T) and P(T) for particular 
i) the effects of changes in R = P /Pj; 
ii) the discontinuous nature of P(T) ; 
iii) the effects of replacing obsolete plant on the energy /year available after investment. 



for 2T t < T <_ (2T 4 + T c ) 
for (2^ + T.) < T <_ 3T £ 
lues of the paramaters, and show 



443 



The parameter values used are T c = 5, T = 1, T = 25, T D = 5 for each plot, with P Q /P i = 3 (Figure 10), P /P. = 2 
(Figure 11) , and P /? t = 1 (Figure 12) . 

Now if one wishes to calculate the total energy output, W(T), up to a time T, one must integrate A. 7, taking 

note of its discontinuities (which can be excluded as sets of measure zero) = For simplicity let (T + T ) < T < 1 

W(T) = -(P N /R) Pe at dt - (P N/R) (1 - e~ aT C) /° +Tp e at dt + (1-£)N P e" a(T c +T P ) / e at dt 
oOO oO rj, ROo T+T 






= -(N P M/aR) + (1 - ^) (P N/a) (e v c P' - 1) (Ao8) 



If T is the time that elapses before the accumulated investment energy is repaid, W(t) = 0, in which case A. 8 

simplifies to 

e aT = e a(T c + T p ) R/(R _ M) (a>9) 

where as before M - e aT P(e aTc - 1) and T = (1/a) In {Re a(T c + V/(r - M) } (A. 10). 

For M < R there is a solution to Equation A. 10, and if M ^ 1 the programme will never repay the energy debt. 
That is, if the programme is ever to repay the debt, P /P. must exceed e aT P(e aTc - 1). 

It must be noted that this is not a sufficient condition because the replacement of obsolete plant must be 
taken into account. This requires integrating A. 7 over more plant lifetimes. To some extent this is a purely 
academic question because for it to arise, the energy required for investment energy/year must exceed or equal 
the annual energy produced from the beginning of production. 



REFERENCES 

1./ Chapman, P F and Mortimer, N, "Energy Inputs and Outputs for Nuclear Power Stations", ERG 005, Energy 
"^ Research Group, Open University, Milton Keynes, Bucks., Sept 1974, revised Dec 1974. 

2. Lovlns, A B, "Nuclear Power: Technical Bases for Ethical Concern", Friends of the Earth Ltd (9 Poland 
St, London W1V 3DG) , 4 Dec 1974. 

3. Lovins, A B, World Energy Strategies: Fasts, Issues, and Options, Friends of the Earth Inc (529 Commercial 
St, San Francisco, California 94111) , early 1975; earlier edition in Bull atom Saient 30, 5, 14-32 (May 
1974) and 30, 6, 38-50 (June 1974); earlier edition (now out of print) published by Earth Resources Research 
Ltd (9 Poland St, London W1V 3DG) , Nov 1973. 

4. CEGB evidence, p 44 in Select Committee on Science and Technology (House of Commons), "The Choice of a 
Reactor System", 73 i-vii, HMSO (London), Feb 1974. 

5. Leslie, D C, p 78 in Inglis, K A D, ed. Energy: From Surplus to Scarcity? , Applied Science Publishers Ltd 
(Ripple Road, Barking, Essex) on behalf of The Institute of Petroleum, Great Britain, 1974. 

6. CEGB, loc dt (ref 4) at p 192. 

7. Searby, P J, Atom 178 (Aug 1971) p 185. 

8. Chapman, P F and Mortimer, N, "Energy Analysis of the Census of Production 1968", ERG 006 (see ref 1), 
in preparation. 

9. Wright, D J, "Calculating Energy Requirements of Commodities from the Input/Output Tables", SARU, Department 
of the Environment (London SW 1) . 

10. Everett, F D, "Mining Practices at Four Uranium Mines", US Bureau of Mines IC-8151. 

11. Bieniewski, C L et al, "Availability of Uranium at Various Prices from Resources in the US", US Bureau of 
Mines IC-8501 (1971). 

12. Clegg, J w and Foley, D D, Uraniur. Ore Processing , Addison-Wesley, 1958. 

13. Anon, "Fuel Cycles for Electrical Power Generation", Teknekron Inc (California), Jan 1973. 

14. OECD, "Uranium: Resources, Production, and Demand", OECD (Paris), Aug 1973. 

15. Anon, "Directory of Power Reactors", Nucl Eng Intl (April 1973). 

16. Anon, "Index of Nuclear Reactors", Nucl Eng Intl (April 1972). 

17. Barnaby, C F, Science J Sh(2) (Aug 1969) , p 54. 

18. Roberts, J T, Intl Atom Energ Bull IS, 5, 14 (Oct 1973). 

19. Moore, J, Bradly, N, and Rowlands, I T, Atom 19S (Jan 1973) p 7. 

20. Data supplied by letter to G Leach from Manager, Heavy Water Pic 

21. Details in "Report on Census of Production 1968" Vol 153 (Table 5) ; see also Chapman, P F in "Conservation 
of Materials", Proc. of conference, Harwell, 26-27 March 1974, p 125. 

22. From estimates given to Chapman by TNPG and UKAEA. 

23. Hill, Sir John, Atom 180 (Oct 1971) p 231. 

24. Bainbrldge, G R and Beveridge, C, Atom 1SS (Sept 1969) p 248. 

25. Bupp, I C et al, "Trends in Light Water Reactor Capital Costs in the United States; Causes and Consequences" 
Center for Policy Alternatives, MIT (Cambridge, Mass), Nov 1974; Techn Rev, forthcoming (about Feb 1975). 



442 



p i N o {(1 

w 



r a Tl) + e -2aT i}e aT 
" aT * + e " 2aTi )e aT 



for 2T l 
for (2T. 



: T <_ (2T + T )-|lst, 2nd, & 

+ T ) < T < 3T ? 3rd 9 enera " 

V T - J V tions 
(A. 5 continued) 



and so on. From A. 4 and A. 5 the investment energy/year as a fraction of output energy/year is 



X. (T) 

1 

X Q (T) 


H 

R 




(1/RXM ♦ e _a(T * " T c " V) 




(M/R) (1 + e" aTt ) 




(1/R){M(1 + e ~ aT h + e" a(2T * 




(M/R) (1 ♦ e" aT * ♦ e" 2aT *) 



for T £ < T <_ (T t 
for (T„ + T_) < 1 



for 2T £ 
for (2T. 



3T„ 



aT n 



aT c 



where M = e" Pfe"* ' - 1) and R = P^P^ (Note that M is independent of time.) So that within each time period 
the fraction of the energy output/year that is required for construction of new plant is also independent of time. 
This fraction increases with the replacement of plant. In Table A.l the values of e _a *, e~ a l , etc are listed 
for a selection of exponential doubling times T and for both T = 25 and T = 50 years. 



T D (years) 




2 




4 




s 




10 




20 




T £ (years) 


25 


50 


25 


50 


25 


50 


25 


50 


25 


50 


-aT£ 
e -2aTfc 
e " 3a T £ 

TABLE A.l. 






VALUES OF e~ 


aT £( 


O 



-2 

e 


0.0133 
0.0002 
0. 
aT if e -3a 


0.0002 

0. 

0. 

r * FOR 


0.0313 
0.0010 

J 0. 

DIFFERED 


0.0010 

0. 

0. 

r doubl: 


0.1770 
0.0314 
0.0055 
NG TIMEi 


0.0314 0.4200 0.1764 
0.0010 0.1764 0.0311 
0.0002 0.0740 0.0055 

AND FOR T = 25, 50 years. 



Analogous values are obtained for higher generations. (The table shows values less than 0.0001 as zeroes.) 



Clearly from Table A.l the effect of replacing obsolete plant is most important for prog 



with a long T 



Consider a programme with T = 5 years. If the plant lifetime is 25 years, then the fraction of annual energy output 
required for investment is 



a(T c 



M/R 

(1/R) (M + 0.133e" 

1.0133M/R 

(1/R) (1.0133M + 0.0002e a(T c + V) 

1.0135M/R 

But for T = 20 and T = 25 the equivalent fractions are 



for (T c + T p > < 
for T t < T < (T^ 
for (T + T ) < 
for 2T < T <_ (2 
for (2T„ + T ) < 



IT, 

T c ) 

1 2T. 



a(T r 



M/R 

(1/R) (M + 0.4200e" 

1.4200M/R 

(1/R) (1.4200M + 0.1764e a 

1.5964M/R 



+ T p ) 



for 


(T c 


+ Tp , 


< T < 


T £ 


tot 


T £ « 


T <_ 


(T t ♦ 


T c ) 


for 


(T * 


+ V 


< T < 


2T * 


for 


2T l 


< T < 


(2T £ 


t T c 


for 


,2T 


+ T 


) < T 





and so on. If we now define P(T) as the annual energy output at time T after subtracting that needed for 
investment, then 



" < P oV R)e 

- (P o N Q /R)(l - e" aT C)e aT 

(1 - (M/R))P o N o e" a(T = + Ve* 


{(1 


- (M/R»e" a(Tc + T P ) - (1 


{(1 


■ (M/R) (1 


+ e -aT t))e -a(T c 


{(1 


- (M/R) (1 
(1/R) (e" 


+ e -*T i)e -a(T c 

2aT c , ■,_ .. aT 
i)}P o N o e 


(1 


- (M/R) (1 


-aTj -2al 



°" P o N e 



l(T„ 



'P'e 



aT 



for 


< 


T <^ T 


c 




for 


T_ < T <_ 


(T c + 


V 


for 


(T c 


♦ v 


< T 


l? t 


for 


T,<Ti 


(T, ♦ 


T c ) 


for 


(T £ 


+ v 


< T 


c 2T, 


for 


2T £ 


< T < 


(2T 


+ T 


for 


CT 


+ T n 


) < T 


< 21 



and so on. Figures 10, 11, and 12 are plots of X (T) and P(T) for particular values of the paramaters, and show 
i) the effects of changes in R = Pq/P^ - 
ii) the discontinuous nature of P(T); 
iii) the effects of replacing obsolete plant on the energy /year available after investment. 



443 



The parameter values used are T c = 5, T = 1, T = 25, T D = 5 for each plot, with P Q /P i = 3 (Figure 10) , ? /?^ = 2 
(Figure 11), and ? /* t = 1 (Figure 12). 

Now if one wishes to calculate the total energy output, W(T), up to a time T, one must integrate A. 7, taking 
note of its discontinuities (which can be excluded as sets of measure zero) . For simplicity let (T + T ) < T < 1 

W(T) = -(P N /R) / C e at dt - (P N /R) (1 - e" aT C) /° +Tp e dt dt + (1-;)NP e" 3 (T C +T p> / e 3t dt 
o oO rp ROo T+T 

= -(N Q P o M/aR) + (1 " f) (P Q N/a) (e a(T " Tc " T p' - 1) (A. 8) 

If T is the time that elapses before the accumulated investment energy is repaid, W(x) = 0, in which case A. 8 
simplifies to 

e *T _ e a(T c + Vr/(r - M) (A. 9) 

where as before M = e aT P(e aTc - 1) and t = (1/a) In {Re a(T c + V/fR - m) } (A. 10). 

For M < R there is a solution to Equation A. 10, and if II > 1 the programme will never repay the energy debt. 
That is, if the programme is ever to repay the debt, P /P. must exceed e P(e c - 1). 

It must be noted that this is not a sufficient condition because the replacement of obsolete plant must be 
taken into account. This requires integrating A. 7 over more plant lifetimes. To some extent this is a purely 
academic question because for it to arise, the energy required for investment energy/year must exceed or equal 
the annual energy produced from the beginning of production. 



REFERENCES 

Chapman, P F and Mortimer, N, "Energy Inputs and Outputs for Nuclear Power Stations", ERG 005, Energy 
Research Group, Open University, Milton Keynes, Bucks., Sept 1974, revised Dec 1974. 

Lovins, A B, "Nuclear Power: Technical Bases for Ethical Concern", Friends of the Earth Ltd (9 Poland 
St, London W1V 3DG) , 4 Dec 1974. 

Lovins, A B, World Energy Strategies: Facts, Issues, and Options, Friends of the Earth Inc (529 Commercial 
St, San Francisco, California 94111), early 1975; earlier edition in Bull atom Scient 30, 5, 14-32 (May 
1974) and 30, 6, 38-50 (June 1974) ; earlier edition (now out of print) published by Earth Resources Research 
Ltd (9 Poland St, London W1V 3DG) , Nov 1973. 

4. CEGB evidence, p 44 in Select Committee on Science and Technology (House of Commons) , "The Choice of a 
Reactor System", 73 i-vii, HMSO (London), Feb 1974. 

5. Leslie, D C, p 78 in Inglis, K A D, ed. Energy: From Surplus to Scarcity?, Applied Science Publishers Ltd 
(Ripple Road, Barking, Essex) on behalf of The Institute of Petroleum, Great Britain, 1974. 

6. CEGB, loc ait (ref 4) at p 192. 

7. Searby, P J, Atom 178 (Aug 1971) p 185. 

8. Chapman, P F and Mortimer, N, "Energy Analysis of the Census of Production 1968", ERG 006 (see ref 1), 
in preparation. 

9. Wright, D J, "Calculating Energy Requirements of Commodities from the Input/Output Tables", SARU, Department 
- of the Environment (London SW 1) . 

10. Everett, F D, "Mining Practices at Four Uranium Mines", US Bureau of Mines IC-8151. 

11. Bieniewski, C L et al, "Availability of Uranium at Various Prices from Resources in the US", US Bureau of 
Mines IC-8501 (1971). 

12. Clegg, J W and Foley, D D, Vraniur. Ore Processing , Addison-Wesley, 1958. 

13. Anon, "Fuel Cycles for Electrical Power Generation", Teknekron Inc (California), Jan 1973. 

14. OECD, "Uranium: Resources, Production, and Demand", OECD (Paris), Aug 1973. 

15. Anon, "Directory of Power Reactors", Nucl Eng Intl (April 1973). 

16. Anon, "Index of Nuclear Reactors", Hucl Eng Intl (April 1972). 

17. Barnaby, C F, Science J SA(2) (Aug 1969) , p 54. 

18. Roberts, J T, Intl Atom Energ Bull IS, 5, 14 (Oct 1973). 

19. Moore, J, Bradly, N, and Rowlands, I T, Atom 195 (Jan 1973) p 7. 

20. Data supplied by letter to G Leach from Manager, Heavy Water Plar 

21. Details in "Report on Census of Production 1968" Vol 153 (Table 5); see also Chapman, P F in "Conservation 
of Materials", Proc. of conference, Harwell, 26-27 March 1974, p 125. 

22. From estimates given to Chapman by TNPG and UKAEA. 

23. Hill, Sir John, Atom 180 (Oct 1971) p 231. 

24. Bainbridge, G R and Beveridge, C, Atom 255 (Sept 1969) p 248. 

25. Bupp, I C et al, "Trends in Light Water Reactor Capital Costs in the United States; Causes and Consequences" 
Center for Policy Alternatives, MIT (Cambridge, Mass), Nov 1974; Techn Rev, forthcoming (about Feb 1975). 



444 



26. Comey, D D, Bull Atom Scient 30, 9, 23 (Nov 1974). 

C27i\ Lem, P N, Odum, H T, and Bolch, W E, "Some Considerations that Affect the Net Yield from Nuclear Power", 
— J paper presented to 19th Annual Meeting, Heal th Physics Society, Houston, Texas, 7-11 July 1974. 

28. USAEC, "Nuclear Power Growth 1974-2000", WASH-0139T7?rrr'Feb 1974, at p 23. 

29. Department of Trade and Industry / Government Statistical Service, "United Kingdom Energy Statistics 1973" 
HMSO (London) , 1974. 

30. USAEC, "Environmental Statement: Management of Commercial High Level and Transuranium-Contaminated Radio- 
active Waste", WASH-1539 (DRAFT), Sept 1974, at pp 3.1-3, 3.1-10, 3.1-11. 

31. Weinberg, A M, Science 177:21 (1972) at p 34. 

32. Fells, I et al, "Energy for the Future", The Institute of Fuel (London), 1973. 



445 



John H Price took his PhD in solid-state physics from Monash 
University (Australia) in 1971, then worked for a time with 
the Australian Commonwealth Scientific and Industrial Research 
Organization. He is currently an energy consultant to Friends 
of the Earth Ltd (9 Poland St, London W1V 3DG, England) , He 
has published a number of technical papers. 



ACKNOWLEDGEMENTS 

Dr Peter Chapman and his colleague Nigel Mortimer provided the 
stimulus for this research with their often-cited paper ERG 005, 
brought it up-to-date with valuable new data, and generously 
provided essential help, advice, and encouragement throughout. 
Dr Chapman also wrote the original draft of §5. Amory Lovins 
edited the text, drafted §4 and §6, and gave indispensable 
technical advice and moral support. The many contributions of 
these three virtual co-authors are acknowledged with gratitude. 

Gerald Leach, another leading practitioner of energy analysis, 
raised some important theoretical questions which he and others 
will (the author hopes) pursue in detail as part of the next 
phase of research into dynamic energy analysis. Walter C Patteron 
also kindly offered technical comments and other help on the MS, 
and, with Colin Blythe, Richard Sandbrook, and Graham Searle, 
commented usefully on matters of style. 

Thanks are also due to various persons within the nuclear community 
for their cooperation in locating obscure data, and to Jane Price, 
Richard Sandbrook, Amory Lovins, and the FOE Ltd office staff for 
their hard work on production and distribution. 

All facts and opinions stated in this paper remain the sole responsi- 
bility of the author, who would be glad to receive any comments 
that might help to improve the MS. 



446 



llii J-J INTERNATIONAL INSTITUTE FOR ENVIRONMENT AND DEVELOPMENT 
President: Barbara Ward (Lady Jackson, D.B.E.) 

London Office 
1525 New Hampshire Ave. N.W. 27 Mortimer Street. 

Washington. D.C. 20036 London. Wl A 4QW 

<202) 462-0900 01-580 7656-7 



NUCLEAR ENERGY BALANCES 
IN A WORLD WITH CEILINGS 



Gerald Leach 



This preliminary paper was prepared for the 
Institute to review nuclear energy balances in 
the context of real world energy needs. It is the 
first report of the Institute's energy accounting 
project now being conducted by the author. The 
Institute welcomes comments on this report. 



All rights reserved, the International Institute 
for Environment and Development 



447 

NUCLEAR ENERGY BALANC ES 
IN A WORLD WITH CEILINGS 



Gerald Leach 



18 December 1974 

International Institute for Environment and Development 
27 Mortimer Street, London, W.l. 



448 



INTRODUCTION 

Several people have recently examined nuclear power from 
the viewpoint of energy input and output talances. In doing 
so they have undermined a very widely helc idea: namely, that 
nuclear power is an almost magical cure fcr fossil fuel 
shortages and tha:, consequently, rapid and indefinitely 
prolonged nuclear growth is a possible, desirable and inevitable 
option. 

Briefly, the argument goes as f ollows . When one totals 
the direct and indirect energy inputs to en entire nuclear fuel/ 
reactor system, two striking facts emerge. First, even if one 
counts only the inputs of hydrocarbon fuels such as coal, oil 
and natural gas.- that is, one ignores fissile fuels such as 
uranium - the inputs form a fairly substantial fraction of the 
energy that will oe delivered by the reactor over its lifetime. 
Second, most of these inputs are compressed into the few years 
of station construction, fuel loading and so forth, before -the 
reactor can begin to deliver any electric power. 

Due to these factors, with present designs and uranium ore 

grades it can take 10-15 years before a nuclear power programme 

with a linear growth of output provides its first gift of net 

energy output. With a sustained exponential growth the situation 

becomes far worse. The programme can become a permanent and 

increasingly large energy sink, absorbing ever more energy as 

it grows, and thus creating an Alice in Wonderland situation. 

in which one runs faster and faster in order to move backwards 

at a rapidly increasing pace. 

c.'V ri'« j A 



449 



Since many nuclear or energy agencies - Including those 
of the UK, France, the USA and the EEC - repeatedly publish 
forecasts for nuclear power in which growth rates are so high 
and/or sustained that they bring the programme well within this 
energy-absorbing Wonderland, these pioneer studies, though 
preliminary, are clearly of the greatest importance. 

This short paper ^ ' takes the topic £. stage further by 
asking a simple question. What happens to nuclear energy 
balances in the context of the real world - a world where the 
needs for energy are not infinite and where' nuclear programmes 
which double in size every two to three years for 40 to 50 years 
have their own Wonderland qualities of mindless extrapolation 
or of wilful and mischievous manipulation of public gullibility? 

In this world of finite needs, nucleai \ e nergy balances 
become spectacularly favourable. Go long as one assumes 
nuclear growth within an overall ceiling for total energy needs, 
the nuclear growth must (a) end sometime, and (b) replace other 
sources of energy and power. Both factors work extravagantly 
hard in favour of the energy balances for nuclear power. Indeed, 
they work so hard that on energetic ground:; only the most rapid 
possible nuclear growth becomes the best strategy for cutting 
consumption of fossil fuels. Paradoxically, the notion that 



(*) This preliminary paper is the first report of a project on 
the energy accounting of energy supply technologies now 
being conducted by the author. The project began on 
1 December 197 ^ and is sponsored by the Inter-national 
Institute for Environment and Development, 27 Mortimer 
Street, London, W.I., England. 



450 



5 - 



nuclear power is an almost magical cure for fuel shortages 
becomes a truism only if one discards the corollary that there- 
fore accelerating growth for long periods is a necessity. 

This paper demonstrates why this should be and briefly 
discusses some awkward policy implications that follow. It is 
very much a preliminary sketch whose purpose is to raise the 
broad issue and to outline urgent questions tha.t need to be 
followed up. Not least of these are longer term problems to 
do with depletion of uranium ores and the rates at which this 
will occur relative; to nuclear growth and the introduction (or 
otherwise) of fuel-breeding technologies. These problems could 
materially alter the conclusions arrived at here, which are 
themselves (we repeat) based only on energy input/output 
considerations. 

TH E BA SI C ARGUM ENT: NUMBERS AND DEFINITIONS 

Before we can look at nuclear energy balances in a real 
world context it is essential to appreciate in greater detail 
the basic methodology and argument of the* 'anti-nuclear case 1 
outlined above. This is contained in two recent papers by 
Peter Chapman and Nigel Mortimer ( Energy Inputs and Outputs for 
Nucle ar Power S t ations ^ ') and by John Price ( Dynamic Energy 
Analysis and Nuclear Power (2)) # Since both papers are based 
on the same input and output numbers (by Chapman and Mortimer), 
were written with close collaboration, and for our purposes here 
reach essentially the same conclusions,- we can treat them 
together. I shall refer to them by an acronym for the authors' 



451 



names: CMP. 

CMP regard a nuclear reactor as a device for converting 
hydrocarbon fuels into electricity: ie, as a direct analogy of 
a conventional oil- or coal-burning power station. Fissile 
materials such as uranium are counted as en energy 'free good'. 

The first stage of the argument is to derive reasonably 
accurate numbers Tor the major energy inputs. to a variety of. 
reactor types. These inputs are the totcl (direct and indirect) 
energy requirements for tasks such as mining and milling 
uranium ores; ur.?jiium concentration and enrichment; fuel rod 
fabrication; hea/y water production; prevision of all materials 
for and construction of the reactor and associated electrical 
equipment etc., etc. and include the energy required to provide 
the fuels and pow::r used in these tasks. They are thus a 
measure of the quantities of fossil fuel energy 'in the ground' 
required to build and run the reactor. Estimates are also 
made of the effective outputs of nuclear stations, allowing for 
typical load factors, distribution losses, and use of power by 
the power stations, offices, showrooms anc so forth of the 
electricity supply system. These outputs arc measures of 
electrical energy actually delivered to final consumers. 

To these data CMP apply several conventions and assumptions: 

1. All nuclear stations take 5 years to construct and 
load with fuel. Energy inputs during this period are evenly 
spread (ic, are at a constant rate) and the total is termed the 
'energy investment', or E^. One can also think - as in all 
the Figures in this paper - in terms of the rate of energy 



452 



investment; that Is, of energy/year, or power (P-i). Since 
Ej is spent over 5 years, E i = 5 x P^. 

2. After completion there is a one year commissioning 
period with no energy/power inputs or outputs. 

3. All stations have a working life of 25 years. During 
this period there are- further energy/power inputs (for refuelling 
etc.) but these are charged against the. station, as it were on 
current account, and are deducted from the outputs of electrical 
power. The result is to give a net outpi.t of power (P ) or, 
over the station lifetime, of energy (E ). Power outputs are 
measured on a continuous basis. Thus for the standard 1000 MW 
station considered, for all types of reactor and with present 
ore grades of about 0.5^ uranium-, P falls; between 500 and 540 MW 
of continuous power. . 

4. The 'energy performance* of a reactor is measured by a 
simple Energy Ratio, E r , defined as total output over total 
input, or Eq/E^ One can also think of a Power Ratio, 

p r = Pq/Pjl. Given the assumptions about construction time and 
working lifetime, E r is always five times P r . 

Now for some results. Using these assumptions, Chapman 
and Mortimer found that, with 0.j$ uranium ore grades, all nine 
reactor types they examined have Energy Ratios in the approxi- 
mate range 10 to 16. The worst performer was the Advanced Gas 
Cooled Reactor (AGR) with 10. 5. The best was the Pressurised 
Water Reactor (PWR) using favourable estimates of annual fuel 
requirements, with a score of 16.5. CMP stress that these 
numbers are approximate but almost certainly on the high side, 



453 



- 6 



since several major energy inputs have been excluded owing to 
lack of available data - including those for waste storage (in 
some cases for 1CP - 1()5 years or more). 

With 0.007^ uranium ores (as. found in the massive 
Chattanooga Shale deposits) Energy Ratios are more widely 
spaced, ranging from 1.28 for Magnox reactors to 6.24 for CANDU, 
These numbers must be considered as very approximate since they 
are extremely sensitive to the energy costs of mining uranium, 
for v/hich the Chapnan and Mortimer estimates are at best 
provisional . 

Throughout the main part of this papei* I shall assume a 
'worst' performance for present ore grades: namely reactor 
programmes based on an Energy Ratio of 10. 

THE BASIC ARGUM ENT 

The first majDr point to emerge from these data is that 
nuclear reactors are extremely efficient energy converters, as 
CMP take pains to point out. An Energy Ratio of 10 means that 
over the reactor lifetime 10 kWh of electricity are delivered 
for every kWh of fossil fuel consumed. The comparable figure 
for conventional coal- or oil-fired stations is about 0.25 k\7h 
delivered - a J tO-fold difference (3). For the UK coal and oil 
extraction-processing-delivery systems the comparable figures 
are close to O.96 and 0.88 kWh delivered for each kWh consumed, 
though here the delivered output is a solid or liquid fuel and 
not electric power w/ . 

Furthermore, once a nuclear station is running the 

'' • ■ aa 



454 



- 7 



performance gap widens markedly. If we Ignore the investment 
to build the station and compare the energy actually delivered 
to the energy used in the form of fossil fuels for refuelling 
etc., the output/ input ratios for nuclear reactors range from 
about 20 to 170 for ,y$ ore grades and f 1 0111 about 2 to 4 for 
0.007/4 ores. Th-B equivalent figures for conventional stations, 
and for coal and oil fuels, remain much as before (namely 0.25, 
O.96 and 0.88) since the energy invested in building these 
systems' is very small compared to the annual fossil fuel inputs 
when the systems are running. 

However, the main plank of the basic argument is not that 
nuclear energy ratios are low but that do* ct ratios (P /Pj.) are 
relatively poor, while the major inputs (l'i) occur before the 
outputs can begin. It is this combination which apparently 
makes nuclear powsr such a doubtful proposition when one 
considers not a single reactor but a growing multi-reactor 
programme . 

FIGURE 1 illustrates the situation for linear growth , where 
work starts in Year 1 to add 16 reactors of 1000 MW capacity 
(and E r = 10) at the rate of one per year. For comparability 
with later Figures, it is assumed that in Year two reactors 
from a previous programme are already working, giving a 
'starting output' of 1000 continuous MW. (Note that with all 
curves of power against time, energy outputs and inputs are 
given. by the area under the curve). 



455 



8 - 



FIGURE 1. TOTAL POWER OUTPUT FOR LIKEAR GROWTH NUCLEAR 
PROGRAMME 




first not energy gain 
from programme 

At the start of the programme the Power Investment (Pj_) 
increases, causing the Total Pov/cr Output to drop. This 
produces a net energy deficit, which is not 'paid back' until 
the few years before the growth ceases. As a result, Total 
Power Output docs not recover to its starting level until 
Year 8, while the first net energy gain from the programme is 
delayed until Year 13>. 

FIGURE 2 shows the situation with a sustaj n od ex pone ntial 
growth programme where the doubling time of the Power Output is 
2 and 4 years (T^ = 2 and l \) . These growth rates are roughly 
comparable to many existing national programmes, yet the Power 
Inputs increase so rapidly that with the best case (T d = ; 
the Total Power Output hardly rises for 30 years and in the 
worst case (T^ = 2) the programme becomes a massive power and 
energy drain. ' '• 



456 



FIGURE 2. POWER AND TOTAL POWER OUTPUTS FOR SUSTAINED 
EXPONENTIAL GROWTH NUCLEAR PROGRAMME . ' 

POWER 

(energy/yr.) 

80i 



60 . POWER OUT (P ) 



40 



20 




POWER 
10i 



20 



•60 




TOTAL POWER OUT (P,,-Pj) 



T d = 4 



The basic argument of CMP is that, while nuclear power Is 
an effective way of turning fossil fuel energy into electrical 
energy, even with linear growth it may not be the best means 
of overcoming immediate fossil fuel shortages and dependencies, 
while with current plans for very rapid and long-sustained 
exponential growths, nuclear power is clearly a complete write-off 



457 



10 - 



Having stated this view, we can now turn to the energetics 
of nuclear power in the context of the real world. We will 
first deal briefly with linear growth to illustrate the main 
new factors which we must consider. 

LINEAR GROWTH WITH I N CEILINGS 

The first question to ask is whether the early power and 
energy deficits are serious in a national energy context, taking 
the UK as a fairly typical example of industrial nations. 

In the UK total primary energy consumption, end use or 
available energy, total electricity output, and nuclear electricity 
output stand in the proportions (1972) of 5.40: 100: 12: 1.7 or 
82: 60: .7:1 ^' . These ratios are for power, or energy/year. 
Using the. latter ' ratios, we can see from Figure 1 that if the 
UK attempted the linear growth programme described, the initial 
power deficit would peak at 1.25 units or about l. l j% of primary 
energy input. This does not seem a large deficit to incur 
in an attempt to boost nuclear output from 1 to 9 units and total 
electricity from 8 to 16 units - that is by 100^. (One should 
also bear in mind that many of the power inputs - eg for mining 
and milling uranium ores - are not incurred in the UK, so that 
the actual fraction would be even smaller.) 

But this is not the whole story. A substantial nuclear 
growth of this kind would almost certainly replace conventional 
power stations and, through a general increase of electrification, 
direct usage of coal, oil, natural gas etc. Given that nuclear 
power is an extremely efficient energy converter in the terms 
used here and. by CMP, we might 'expect some .rather dramatic 



458 



li 



consequences when- we impose ceilings on total energy usage. 

Haw dramatic these consequences could theoretically be, 
we show in FIGURE '$, where we take the highly artificial 
situation of a static electricity output within which nuclear 
grows linearly (with the same conditions of Figure 1) from 10 
to 90;^ of output. 

The Total Power input for electricity production falls 
dramatically. There is a small energy penalty; to pay (the 
black area on the 1 ovzer paraph) but the eventual energy savings 
ie, savings of fossil fuels - are enormous: they are the area 
between the full ar d dotted lines. 



FIGURE ?. FOSSIL FUEL REQUIREMENTS FOR LINEAR NUCLEAR GROWTH 
WITHIN / FIXED CEILING FOR ELECTRICITY PRODUCTION 



POWER 
(energy/yr.) 



10 



10 


TOTAl ELECT 


1IC POWER 


OUT 




nor. -nuclear 






nucloar 





, . 


r 1 



20 



years 



40 



20- 



POWER IN FOR 
TOTAL ELECTRIC 
OUTPUT 



without nuclear 
growth 




10 



20 



years 



459 



It is intuitively obvious, and not hard to prove, that 
for any given ceiling for total electricity output (whether 
rising, constant or declining) the largest fuel savings occur 
with the largest rates of nuclear increase. With such rapid 
nuclear growth the initial power or energy deficit for nuclear 
increases, but this, deficit is also 'paid tack' sooner when the 
nuclear growth reaches- its upper limits and flattens off (see 
Figure l). To cut fuel consumption, the best strategy is 
therefore to go for the fastest nuclear growth that gives an 
initial deficit one can just 'afford'. However, as we shall 
discuss later, very rapid nuclear growth rates - and sudden 
changes in these rotes - produce their own problems so that the 
actual trade-offs to be made are more complicated than this 
simple model suggests. - 



460 



15 



EXPONENTIA L GROWTH W I TH CEI L INGS 

Before we look at the more likely case of exponential 
nuclear growth we need to consider the actual energy growth 
conditions of industrial societies and the absurdities produced 
by very rapid and sustained exponentials. 

In the UK the "amount of energy per year that is available 
to consumers - that is, the fuel that goes into the tank, the 
coal shovelled into the industrial boiler, the electricity that 
one switches on, and so forth - has increased by only ^0% or so 
since 1900 ^5), This, 'end use' or available energy increased 
from 196O-70 by only 1.358$ per year, while from 1969-72 the rise 
was a mere 0.75$ per annum ( ~'' . Available energy has been 
rising more rapidly in the USA, with a 3.9$ p. a. average increase 
from^ 1957-71, slowing slightly to 3.&/0 p. a. from 1968-71 ' K 

In both countries primary energy use has, of course, been 
rising considerably faster. It is easy to see why. This 
primary measure includes the energy used by the energy supply 
industries in providing available energy anc recent decades have 
seen ma.ssive switches from coal to oil and from solid or liquid 
fuels to electricity - both of them trends which raise (and with 
conventional electricity production, raise enormously) the 
primary energy requirements for 'end use' or available energy. 
At the same time, a switch from manufactured to natural gas- 
offsets these trends, which helps to explain why in the USA the 
growth of primary energy has been little higher than that of 
available energy. 



461 



- 14 - 



Be that as it may, it is clearly the 'end use' energy 
which should be the basic measure that one uses in forecasting 
energy needs. Primary energy, so often predicted to continue 
rising at high historical rates, is merely s secondary measure 
which depends, firs;, on the level of available energy, and 
second, on the mix of technologies used to provide it. 

With these 'end use' growth rates of 1% and l \% p. a. in 

mind, it is instructive to look at FIGURE k . 'Starting at Year 

with 'a UK pattern wiore end use consumption and nuclear output 

are in the proportion 100: 1.7, very rapid and sustained 

exponential nuclear growths soon produce absurd results - even 

when one takes US rates of growth for end use energy and makes 

no allowance at all for their steady reduction through naturally 
7) 



o c c u i • r ]. n g c > 1 ;. xn; o s 



or deliberate conservation measures. 



FIGURE 4. RAPID AND SUSTAINED NUCLEAR GROW'. 'H IN THE CONTEXT 
OF TOTAL ENERGY GROWTH 



POWER 

(ei.er S y/year) 

SOO- 



400 



300 



200 



100 



NUCLEAR OUTPUT 



/4/. 



TOTAL ENERGY 
CONSUMPTION 

(end use bc'isis) 







2-5 7. 



462 



15 



Under these conditions, or any remotely like them, very- 
rapid nuclear exponentials cannot be sustained for very long; 
and the more rapid the growth, the sooner it must end. 

EFFECTS OF ENDING EXPONENTIAL NUCL EA R GROWTH 

Ending a nuclear exponential growth programme has a 
remarkable impact on its energy or power balances. This is 
because during a sustained exponential growth (using the fore- 
going assumptions) the Pov/er Input in any year (k) i s 
proportional to the sum of the annual increases in Power Output 
in Years (t+2), (t-i-;i)> (t+'i) , (t+5) and (t+6). The latter 
increments are of course very much larger than the earlier ones. 
Cutting off or cutting down these increments by suddenly ending 
or slowing down the growth therefore has a dramatic effect on 
the Power Input up to 6 years before the slow-down or cut-off 
point. 

FIGURE 5 shows this effect for two massive nuclear • 
programmes, A and B, which start with rapid exponential growths 
of doubling time 2 end 4 years respectively, then increase 
linearly, and finally approach a plateau at a level of output 
60 times higher than the starting level (l power unit). The 
resulting Total Power Output (P Q - Pj ) Is shown below and should 
be compared with the curves of Figure 2. Instead of becoming 
a vast and increasing pov/er or energy sink, Programme A has an 
initial power or energy deficit but begins to pay it off very 
rapidly after Y'ear 12. Programme B produces a trivially small 
deficit for 21 years, and then begins to pay back at an almost 



463 



- 16 - 



equally rapid pace, 



FIGURE 5. POWER AlfD TOTAL POWER OUTPUTS FOR CURTAILED 
EXPONENTIAL NUCLEAR PROGRAMMES 

POWER 

(en3rgy/yr.) 
60 

POWER OUT (p 




Taking these curves in isolation from the rest of the 
energy system, programme A is clearly the best option (on 
energy balance grounds only ) - provided that the initial power 
or energy deficit is relatively small. 

In fact it is, relatively small considering